[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"blog-posts":3},{"posts":4,"source":295,"total":296},[5,25,39,52,64,76,91,103,114,127,139,151,163,175,188,200,213,226,239,250,263,274,285],{"slug":6,"title":7,"excerpt":8,"content":9,"image":10,"category":11,"author":12,"status":13,"reviewed":14,"views":15,"metaTitle":16,"metaDescription":17,"frameworks":18,"productId":22,"publishedAt":23,"updatedAt":24,"date":23},"fivem-rp-server-new-player-onboarding-first-hour","The First Hour on a FiveM RP Server: Onboarding That Keeps Players","Most new players decide in their first session whether to come back. A practical onboarding plan for FiveM roleplay servers, from Discord to the first paycheck.","\u003Cp>Most FiveM roleplay servers do not lose new players to bad scripts or lag. They lose them in the first hour, when someone spawns at the airport with no money, no idea what to do and nobody to talk to, and quietly disconnects. This article is a practical onboarding plan for server owners, from the Discord invite to the first paycheck, plus a simple way to measure whether it works.\u003C\u002Fp>\u003Ch3>Why the first session matters more than any feature\u003C\u002Fh3>\u003Cp>Veteran players forgive a lot because they already have friends, a job and a reason to log in. A new player has none of that. Everything they see in the first session is your server, and if the honest summary is \"I walked around confused and got run over\", they will not open FiveM again to give you a second chance.\u003C\u002Fp>\u003Cp>The goal of onboarding is not to explain every system. It is to get a new player from \"I just connected\" to \"I have a plan for tomorrow\" before they lose interest. That usually means three things: they understand where they are, they earned something, and they talked to at least one real person.\u003C\u002Fp>\u003Ch3>Before they join: Discord and the connection screen\u003C\u002Fh3>\u003Cp>Onboarding starts before the game loads. Walk through your own funnel with a fresh Discord account and write down every step. Common problems:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Too many channels visible on day one.\u003C\u002Fstrong> A new member sees forty channels and no obvious starting point. Hide most of them behind a role and show a single welcome channel with three links: rules, how to connect, where to ask questions.\u003C\u002Fli>\u003Cli>\u003Cstrong>Whitelist applications with no feedback.\u003C\u002Fstrong> If you run a whitelist, tell applicants how long review takes and actually meet that time. An application that sits for a week is a player you already lost.\u003C\u002Fli>\u003Cli>\u003Cstrong>Rules written as a legal document.\u003C\u002Fstrong> Nobody reads fifty numbered rules. Keep a short version with the ten rules that actually get people banned, and link the full document for reference.\u003C\u002Fli>\u003Cli>\u003Cstrong>No connection instructions.\u003C\u002Fstrong> Spell out how to find the server, what to do if a download stalls, and that the first load takes longer than the next ones.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>The loading screen is the last thing they see before spawning. Use it to set expectations: what kind of city this is, where to go first, and how to ask staff for help in game. One clear instruction beats a slideshow of screenshots.\u003C\u002Fp>\u003Ch3>The first five minutes in game\u003C\u002Fh3>\u003Cp>Character creation is often where new players spend the most time and feel the least guided. Give them sensible defaults and make it obvious which steps are optional. If your clothing menu has hundreds of drawables, suggest a quick outfit and let them come back later.\u003C\u002Fp>\u003Cp>Then think hard about where they spawn. A random airport spawn with no direction is the worst option. Better choices:\u003C\u002Fp>\u003Cul>\u003Cli>A fixed new player spawn near a job center, a cheap clothing store and a bus or taxi stop.\u003C\u002Fli>\u003Cli>A small, quiet area where they will not be in the middle of a police chase within thirty seconds.\u003C\u002Fli>\u003Cli>A visible next step: a marker, a ped to talk to, or a notification that says exactly where to go.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>If your framework gives new characters a phone, make sure the first thing on it is useful: a contact for help, a map pin to the job center, or a note with the server basics.\u003C\u002Fp>\u003Ch3>A first job that teaches the city\u003C\u002Fh3>\u003Cp>The best tutorial is a job that pays. Delivery, garbage collection, bus routes and similar legal jobs work well for new players because they move the player across the map, introduce driving and vehicle rules, and end with money in their pocket.\u003C\u002Fp>\u003Cp>When you choose or configure starter jobs, check them against this list:\u003C\u002Fp>\u003Col>\u003Cli>\u003Cstrong>No upfront cost.\u003C\u002Fstrong> The job should not require a vehicle, license or tool the player cannot afford yet.\u003C\u002Fli>\u003Cli>\u003Cstrong>Short first loop.\u003C\u002Fstrong> The first payout should arrive within ten to fifteen minutes, not after an hour of grinding.\u003C\u002Fli>\u003Cli>\u003Cstrong>Route through key locations.\u003C\u002Fstrong> Send them past the hospital, police station, garages and shops so they learn the city while working.\u003C\u002Fli>\u003Cli>\u003Cstrong>Works solo, better in a group.\u003C\u002Fstrong> A job that can be done alone but pays more with a partner nudges new players toward meeting people.\u003C\u002Fli>\u003Cli>\u003Cstrong>Low risk.\u003C\u002Fstrong> Losing a whole shift of income to a random robbery on day one is a fast way to lose a player.\u003C\u002Fli>\u003C\u002Fol>\u003Ch3>Starter money and items\u003C\u002Fh3>\u003Cp>Starter cash is a balancing act. Too little and new players cannot do anything; too much and veterans make alts to farm it. A practical approach is to give enough for a phone, basic food and a few bus or taxi rides, then let the first job do the rest.\u003C\u002Fp>\u003Cp>Avoid handing out vehicles or weapons in starter packs. They devalue the economy and skip the part of the game where a new player sets a goal and works toward it. A cheap starter car that becomes available after a few hours of play, or after completing a first job, gives them that goal.\u003C\u002Fp>\u003Ch3>Staff and community: someone has to say hello\u003C\u002Fh3>\u003Cp>Scripts cannot replace a real person. The single most effective onboarding change many servers make is a small team of greeters: trusted players or junior staff who keep an eye out for new arrivals and offer help in character.\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Mark new players discreetly.\u003C\u002Fstrong> A short-lived tag visible to staff, or a Discord role for the first week, lets helpers find them without breaking immersion for everyone else.\u003C\u002Fli>\u003Cli>\u003Cstrong>Treat early mistakes as teaching moments.\u003C\u002Fstrong> A new player who breaks a rule they did not know about should get a quiet explanation, not a ban. Save strict enforcement for repeat or obvious bad faith behavior.\u003C\u002Fli>\u003Cli>\u003Cstrong>Brief your criminal community.\u003C\u002Fstrong> Agree as a server that brand new characters are not prime targets for robberies or kidnappings in their first sessions. Most experienced roleplayers will respect this if you explain why.\u003C\u002Fli>\u003Cli>\u003Cstrong>Have one clear help command or channel.\u003C\u002Fstrong> Whatever you use, every new player should know it exists before they spawn.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Measure where players drop off\u003C\u002Fh3>\u003Cp>You cannot fix onboarding by guessing. At minimum, know how long new players stay in their first session and why they disconnect. A small server-side resource can log this:\u003C\u002Fp>\u003Cpre>\u003Ccode>local sessions = {}\n\nAddEventHandler('playerJoining', function()\n    sessions[source] = os.time()\nend)\n\nAddEventHandler('playerDropped', function(reason)\n    local started = sessions[source]\n    if not started then return end\n\n    local license = GetPlayerIdentifierByType(source, 'license')\n    local minutes = math.floor((os.time() - started) \u002F 60)\n\n    print(('[onboarding] %s left after %d min (%s)'):format(license or 'unknown', minutes, reason))\n    sessions[source] = nil\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>In production, write this to your database instead of the console and join it against your players table so you can separate first sessions from returning players. Then look for patterns:\u003C\u002Fp>\u003Cul>\u003Cli>If many new players leave within five minutes, the problem is spawn, character creation or the first instruction.\u003C\u002Fli>\u003Cli>If they leave after twenty to thirty minutes, they probably ran out of things to do or money to do them with.\u003C\u002Fli>\u003Cli>If they finish a good first session but never return, they did not form a connection or a goal.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Numbers only tell you where. Ask a handful of new players directly what confused them. A short message from a staff member after their first day often gets honest, specific answers.\u003C\u002Fp>\u003Ch3>A quick onboarding checklist\u003C\u002Fh3>\u003Col>\u003Cli>Join your own Discord with a fresh account and fix every confusing step.\u003C\u002Fli>\u003Cli>Cut your rules down to a readable short version with a link to the full set.\u003C\u002Fli>\u003Cli>Use the loading screen for one clear instruction, not decoration.\u003C\u002Fli>\u003Cli>Set a fixed, safe new player spawn next to a job center and transport.\u003C\u002Fli>\u003Cli>Configure at least two starter jobs with no upfront cost and a fast first payout.\u003C\u002Fli>\u003Cli>Give starter cash for essentials only; no free vehicles or weapons.\u003C\u002Fli>\u003Cli>Recruit greeters and agree on a grace period for brand new characters.\u003C\u002Fli>\u003Cli>Log first session length and disconnect reasons, and review them weekly.\u003C\u002Fli>\u003C\u002Fol>\u003Ch3>Takeaway\u003C\u002Fh3>\u003Cp>Onboarding is not a single script or a tutorial popup. It is the path from Discord invite to first paycheck, and every confusing step on that path costs you players. Walk it yourself as a stranger, make the first job pay quickly, make sure a real person says hello, and track where people leave. Fix the biggest drop-off first, then measure again.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-rp-server-new-player-onboarding-first-hour.webp?v=mwnukap1","Tutorials","ATY Development","published",false,6,"FiveM RP New Player Onboarding: The First Hour","How to onboard new players on a FiveM roleplay server: Discord flow, spawn, first job, starter money, staff greeters and tracking where players drop off.",[19,20,21],"QBCore","ESX","Qbox",null,"2026-11-19T20:28:26+00:00","2026-11-20T23:33:01.728945+00:00",{"slug":26,"title":27,"excerpt":28,"content":29,"image":30,"category":31,"author":12,"status":13,"reviewed":14,"views":32,"metaTitle":33,"metaDescription":34,"frameworks":35,"productId":36,"publishedAt":37,"updatedAt":38,"date":37},"fivem-loading-screen-v5-real-progress-themes","FiveM Loading Screen V5: Real Load Progress, Three Themes, One Config","A practical look at ATY Loading Screen V5: progress tied to real load events, three separate theme layouts, queue mode and a single config.lua for staff.","\u003Cp>Every player on your server spends time staring at the loading screen, often longer than they spend reading your rules channel. Most servers waste that window with a static image and a progress bar that has nothing to do with what the game is actually doing. ATY Loading Screen V5 turns that wait into something useful: honest progress, real server information and a proper hand-off into the city.\u003C\u002Fp>\u003Ch3>The problem with most loading screens\u003C\u002Fh3>\u003Cp>A typical loading screen fails in three predictable ways. The bar is fake, usually a timer that fills to 90% and then sits there while the client streams assets. The content is stale, because changing a line of text means editing HTML or rebuilding a bundle. And when something goes wrong on the server side, the screen never closes and the player is left looking at a frozen backdrop until they quit.\u003C\u002Fp>\u003Cp>V5 is built around fixing exactly those points. Here is how each piece works in practice.\u003C\u002Fp>\u003Ch3>Three themes with their own layouts\u003C\u002Fh3>\u003Cp>The resource ships three complete themes: \u003Cstrong>cinematic\u003C\u002Fstrong>, \u003Cstrong>cockpit\u003C\u002Fstrong> and \u003Cstrong>minimal\u003C\u002Fstrong>. These are separate layouts, not one design recoloured three ways. That matters because the right layout depends on your server identity:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Cinematic\u003C\u002Fstrong> suits servers that want the screen to feel like an intro sequence before the city.\u003C\u002Fli>\u003Cli>\u003Cstrong>Cockpit\u003C\u002Fstrong> fits communities that like an information-dense, dashboard-style presentation.\u003C\u002Fli>\u003Cli>\u003Cstrong>Minimal\u003C\u002Fstrong> works when you want the backdrop and progress front and centre with little else competing for attention.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>You can lock the theme server-side so every player sees your chosen design, or let players switch between them from their settings. The same config-gated choice applies to language.\u003C\u002Fp>\u003Ch3>Progress that matches what FiveM is doing\u003C\u002Fh3>\u003Cp>The progress display tracks the game's real load events rather than a timer. According to the product description, it listens to 13 native load events, groups them into six weighted phases and shows live sub-progress inside each phase. The number on screen is the number FiveM is actually at.\u003C\u002Fp>\u003Cp>This sounds like a cosmetic detail until you deal with support tickets. When a player says they are stuck at a specific point, an honest bar tells you something. A fake bar that always stalls at the same percentage tells you nothing and trains players to assume the server is broken.\u003C\u002Fp>\u003Ch3>What players actually see\u003C\u002Fh3>\u003Cp>Beyond the progress bar, the screen is built to give the player something relevant to look at while assets stream in.\u003C\u002Fp>\u003Ch4>Personalised welcome\u003C\u002Fh4>\u003Cp>Returning players can be greeted with their name, job, playtime and last login. First-join detection is done via handover, so a brand new player can be treated differently from a regular. Handover is the standard FiveM mechanism for passing data from the server to the loading screen during connection, which is why this can work before the player has spawned.\u003C\u002Fp>\u003Ch4>Discord avatars\u003C\u002Fh4>\u003Cp>Player avatars can be pulled from Discord. A Discord token is optional: without one, the screen falls back to the default avatar instead of breaking or showing an empty frame.\u003C\u002Fp>\u003Ch4>Music player\u003C\u002Fh4>\u003Cp>The music player supports local audio files, with a waveform generated from the audio itself, or a YouTube embed that does not require an API key. Players control volume, shuffle and repeat from their settings, so nobody is forced to listen to your playlist at full volume.\u003C\u002Fp>\u003Ch4>Tabbed content panels\u003C\u002Fh4>\u003Cp>Announcements, Changelog, Rules, Staff, Guide, Keybinds and Gallery live in tabbed panels. They are closed by default and opened from a stacked rail, so the screen stays clean for players who just want to wait, while anyone curious can read your latest changelog or check keybinds. The gallery has a click-through lightbox, and rotating tips use dot navigation and pause on hover.\u003C\u002Fp>\u003Ch4>Player settings\u003C\u002Fh4>\u003Cp>Players can adjust theme, language, background mode, volume, shuffle, repeat, reduced motion and film grain, subject to whatever you lock server-side. Reduced motion is worth highlighting: some players genuinely find heavy animation uncomfortable, and giving them an off switch costs you nothing.\u003C\u002Fp>\u003Ch3>What staff control\u003C\u002Fh3>\u003Cp>The part server owners will care about most is maintenance. Every colour, font, string, panel and asset is driven from a single \u003Ccode>config.lua\u003C\u002Fcode>, and nothing needs a rebuild to change. Updating your rules tab before a wipe, adding a changelog entry after a patch or swapping the backdrop for an event becomes a config edit and a resource restart, not a front-end build step.\u003C\u002Fp>\u003Cp>A few other staff-facing features are worth knowing about:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Queue mode.\u003C\u002Fstrong> When players are waiting in a queue, the screen shows position, queue length and estimated wait as its own state, instead of pretending the game is loading.\u003C\u002Fli>\u003Cli>\u003Cstrong>Safety timeout.\u003C\u002Fstrong> If the server never signals that it is ready, the screen force-shuts so nobody gets stranded on a frozen loading screen.\u003C\u002Fli>\u003Cli>\u003Cstrong>Enter gate and exit choreography.\u003C\u002Fstrong> The screen uses manual shutdown and an animated hand-off into the city, so the transition into gameplay is deliberate rather than an abrupt cut.\u003C\u002Fli>\u003Cli>\u003Cstrong>Localization.\u003C\u002Fstrong> Full TR\u002FEN localization is included. Any missing key falls back to the default locale, so a partially translated string never prints \u003Ccode>undefined\u003C\u002Fcode> on screen.\u003C\u002Fli>\u003Cli>\u003Cstrong>Easter egg.\u003C\u002Fstrong> A hidden Konami-style easter egg is included and gated by config, so you can disable it if it does not fit your server.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>How it fits a QBCore or ESX setup\u003C\u002Fh3>\u003Cp>In FiveM, a loading screen is a regular resource that declares an HTML page as its loadscreen in the manifest. When a resource wants to control when the screen closes instead of letting the game close it automatically, it opts into manual shutdown. A generic example of how FiveM loading screens are declared looks like this:\u003C\u002Fp>\u003Cpre>\u003Ccode>fx_version 'cerulean'\ngame 'gta5'\n\nloadscreen 'html\u002Findex.html'\nloadscreen_manual_shutdown 'yes'\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>That snippet is only there to illustrate the mechanism; the resource ships with its own manifest and you should use that as-is. The relevant FiveM behaviour is documented at \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\" target=\"_blank\" rel=\"noopener noreferrer\">docs.fivem.net\u003C\u002Fa> if you want to understand what is happening under the hood.\u003C\u002Fp>\u003Cp>For a typical QBCore or ESX server, practical integration notes:\u003C\u002Fp>\u003Col>\u003Cli>Remove or stop any existing loading screen resource first. Two resources both declaring a loadscreen is a common source of confusion.\u003C\u002Fli>\u003Cli>Ensure the resource in your \u003Ccode>server.cfg\u003C\u002Fcode> like any other resource.\u003C\u002Fli>\u003Cli>Open \u003Ccode>config.lua\u003C\u002Fcode> and set your theme policy, language policy, panel content and assets before going live.\u003C\u002Fli>\u003Cli>Decide whether you want Discord avatars with a token or are happy with the default avatar fallback.\u003C\u002Fli>\u003Cli>Test a fresh connection and a returning connection so you see both the first-join and returning welcome.\u003C\u002Fli>\u003C\u002Fol>\u003Cp>On performance, the screen is built for CEF, the embedded browser FiveM uses for UI. The description lists a WebP backdrop, pre-rendered blur backing and a working \u003Ccode>backdrop-filter\u003C\u002Fcode>, at around 695 KB total. A lighter loading screen means less for the client to fetch before it can start on the actual game.\u003C\u002Fp>\u003Cp>The resource is fully open source: the complete Vue source is included and nothing is escrowed. If your developer wants to restructure a theme or add a panel beyond what the config exposes, they have the code to do it.\u003C\u002Fp>\u003Ch3>Who it is for\u003C\u002Fh3>\u003Cul>\u003Cli>\u003Cstrong>Roleplay servers with regular updates\u003C\u002Fstrong> that want changelogs, announcements and rules visible at the moment every player connects.\u003C\u002Fli>\u003Cli>\u003Cstrong>Servers that run a queue\u003C\u002Fstrong> and want players to see an honest position and wait estimate.\u003C\u002Fli>\u003Cli>\u003Cstrong>Owners without a front-end developer\u003C\u002Fstrong> who need to change text, colours and assets without touching a build pipeline.\u003C\u002Fli>\u003Cli>\u003Cstrong>Servers with in-house developers\u003C\u002Fstrong> who want full source access rather than an escrowed UI.\u003C\u002Fli>\u003Cli>\u003Cstrong>Turkish and English communities\u003C\u002Fstrong>, given the TR\u002FEN localization out of the box.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>If you are setting up the rest of the connection flow as well, the \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Fproduct\u002F7030079\">Multi Character & Spawn Selector\u003C\u002Fa> covers what happens after the loading screen hands the player off.\u003C\u002Fp>\u003Ch3>Practical takeaway\u003C\u002Fh3>\u003Cp>A loading screen is the first piece of your server every player sees, every single session. Before you pick one, check these points:\u003C\u002Fp>\u003Cul>\u003Cli>Does the progress reflect real load events, or is it a timer?\u003C\u002Fli>\u003Cli>Can staff update rules, announcements and changelogs without a rebuild?\u003C\u002Fli>\u003Cli>Is there a safety timeout so a server hiccup does not strand players?\u003C\u002Fli>\u003Cli>Does it handle queue state separately from loading?\u003C\u002Fli>\u003Cli>Can players reduce motion and control audio?\u003C\u002Fli>\u003Cli>Do you get the source if you need to change something the config does not cover?\u003C\u002Fli>\u003C\u002Ful>\u003Cp>ATY Loading Screen V5 answers yes to all of them. Full details and support contact are on the \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Fproduct\u002F7608535\">product page\u003C\u002Fa>.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-loading-screen-v5-real-progress-themes.webp?v=mwjmwz9g","Scripts",5,"FiveM Loading Screen V5: Real Progress and Three Themes","How ATY Loading Screen V5 works on a QBCore or ESX server: real load progress, three themes, personalised welcome, queue mode and one config.lua.",[19,20],7608535,"2026-11-19T17:32:28.492+00:00","2026-11-20T23:36:01.913965+00:00",{"slug":40,"title":41,"excerpt":42,"content":43,"image":44,"category":11,"author":12,"status":13,"reviewed":14,"views":45,"metaTitle":46,"metaDescription":47,"frameworks":48,"productId":22,"publishedAt":50,"updatedAt":51,"date":50},"fivem-server-cfg-explained","FiveM server.cfg Explained: What Each Line Does and Where It Goes","A line-by-line walkthrough of a FiveM server.cfg: endpoints, convars, license key, database string, ensure order, permissions and how to keep secrets out of it.","\u003Cp>Most broken FiveM servers are not broken by a script. They are broken by a \u003Ccode>server.cfg\u003C\u002Fcode> that grew by copy and paste until nobody knew which line did what. This guide walks through the file section by section, explains why order matters, and shows a layout that stays readable as your resource list grows.\u003C\u002Fp>\u003Ch3>What server.cfg actually is\u003C\u002Fh3>\u003Cp>The config file is not a settings file in the usual sense. It is a list of console commands that the server runs in order at startup, exactly as if you had typed them into the server console. That has two consequences:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Order matters.\u003C\u002Fstrong> A resource started on line 40 cannot rely on something that only starts on line 80.\u003C\u002Fli>\u003Cli>\u003Cstrong>Anything you can type in the console can go in the file.\u003C\u002Fstrong> That includes \u003Ccode>exec\u003C\u002Fcode>, which loads another config file, so you can split one long file into several smaller ones.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>If you use txAdmin, it still reads a \u003Ccode>server.cfg\u003C\u002Fcode>; txAdmin just gives you an editor for it and handles starting and restarting the process. The rules below apply either way. The official reference for commands is the \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\u002Fdocs\u002Fserver-manual\u002Fserver-commands\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">server commands page\u003C\u002Fa> on the FiveM docs.\u003C\u002Fp>\u003Ch3>Network endpoints\u003C\u002Fh3>\u003Cp>The first lines usually define where the server listens:\u003C\u002Fp>\u003Cpre>\u003Ccode>endpoint_add_tcp \"0.0.0.0:30120\"\nendpoint_add_udp \"0.0.0.0:30120\"\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>\u003Ccode>0.0.0.0\u003C\u002Fcode> means listen on all interfaces. 30120 is the default port. If you run several servers on one machine, give each a different port and open both TCP and UDP for it in your firewall. A server that shows up in the list but cannot be joined is very often a firewall that only allows one of the two protocols.\u003C\u002Fp>\u003Ch3>set, sets and setr: the three kinds of convar\u003C\u002Fh3>\u003Cp>Convars are named variables that the server and resources read. How you define them decides who can see them:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ccode>set\u003C\u002Fcode> defines a convar on the server only. Use it for anything secret, such as database credentials.\u003C\u002Fli>\u003Cli>\u003Ccode>sets\u003C\u002Fcode> defines a convar and publishes it in the server info, so it is visible in the server list and to anyone querying your server. Use it for tags, locale and descriptions.\u003C\u002Fli>\u003Cli>\u003Ccode>setr\u003C\u002Fcode> defines a convar and replicates it to clients, so client scripts can read it. Useful for client-side options a resource expects, never for secrets.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>A common mistake is writing \u003Ccode>setr\u003C\u002Fcode> or \u003Ccode>sets\u003C\u002Fcode> in front of a connection string because a tutorial did. If a value contains a password, it should be \u003Ccode>set\u003C\u002Fcode> and nothing else.\u003C\u002Fp>\u003Ch3>Server identity and listing\u003C\u002Fh3>\u003Cpre>\u003Ccode>sv_hostname \"My City RP | Serious RP\"\nsets sv_projectName \"My City RP\"\nsets sv_projectDesc \"Whitelisted serious roleplay\"\nsets tags \"roleplay, serious, economy\"\nsets locale \"en-US\"\nload_server_icon logo.png\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>The hostname and project name are what players see in the browser. Keep tags honest and short; players filter by them. The server icon must be a small square PNG placed next to your config file. If you want a private development server that does not appear in the public list, the usual approach is to blank the master server convar with \u003Ccode>sv_master1 \"\"\u003C\u002Fcode>.\u003C\u002Fp>\u003Ch3>License key, slots and game build\u003C\u002Fh3>\u003Cpre>\u003Ccode>sv_licenseKey \"your_key_here\"\nsv_maxclients 48\nset sv_enforceGameBuild &lt;build&gt;\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>The license key comes from the Cfx.re portal and ties the server to your account. Treat it like a password: anyone with it can run a server under your name.\u003C\u002Fp>\u003Cp>\u003Ccode>sv_maxclients\u003C\u002Fcode> sets the slot count. Higher slot counts require OneSync to be enabled, which txAdmin exposes as a setting and which can also be set on the command line. Check the current limits for your tier on the docs rather than trusting an old forum post.\u003C\u002Fp>\u003Cp>\u003Ccode>sv_enforceGameBuild\u003C\u002Fcode> forces clients onto a specific GTA game build. Many vehicles, clothing and map packs depend on a minimum build, so if a resource description says it needs a newer build, this is the line to change. Changing it forces every player to download that build, so do it deliberately and announce it.\u003C\u002Fp>\u003Ch3>Security-related lines\u003C\u002Fh3>\u003Cpre>\u003Ccode>sv_scriptHookAllowed 0\nsv_endpointprivacy true\u003C\u002Fcode>\u003C\u002Fpre>\u003Cul>\u003Cli>\u003Ccode>sv_scriptHookAllowed 0\u003C\u002Fcode> blocks clients from using ScriptHook-based mods. Leave it off on any public server.\u003C\u002Fli>\u003Cli>\u003Ccode>sv_endpointprivacy true\u003C\u002Fcode> hides player IP addresses from other players in the player list data.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Avoid setting an \u003Ccode>rcon_password\u003C\u002Fcode> unless you have a specific tool that needs it. txAdmin and the live console cover what most owners used RCON for, and an exposed RCON is one more thing to attack.\u003C\u002Fp>\u003Ch3>The database connection string\u003C\u002Fh3>\u003Cp>Most QBCore, Qbox and current ESX setups use oxmysql, which reads a convar for its connection:\u003C\u002Fp>\u003Cpre>\u003Ccode>set mysql_connection_string \"mysql:\u002F\u002Fdbuser:dbpass@127.0.0.1\u002Fmy_city?charset=utf8mb4\"\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Points worth checking when the database refuses to connect:\u003C\u002Fp>\u003Cul>\u003Cli>The database name in the string must already exist; oxmysql will not create it.\u003C\u002Fli>\u003Cli>Special characters in the password can break the URL format. Either use a password without \u003Ccode>@\u003C\u002Fcode>, \u003Ccode>:\u003C\u002Fcode> or \u003Ccode>\u002F\u003C\u002Fcode>, or URL-encode it.\u003C\u002Fli>\u003Cli>Use \u003Ccode>127.0.0.1\u003C\u002Fcode> if the database is on the same machine. A remote database adds latency to every query.\u003C\u002Fli>\u003Cli>Create a dedicated database user for the server. Do not use the root account.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Starting resources: ensure, start and categories\u003C\u002Fh3>\u003Cp>\u003Ccode>ensure name\u003C\u002Fcode> starts a resource, or restarts it if it is already running. \u003Ccode>start\u003C\u002Fcode> only starts it. In practice, use \u003Ccode>ensure\u003C\u002Fcode> everywhere.\u003C\u002Fp>\u003Cp>Folders in square brackets, such as \u003Ccode>resources\u002F[qb]\u003C\u002Fcode>, are categories. The server searches inside them for resources, and you can ensure a whole category with \u003Ccode>ensure [qb]\u003C\u002Fcode>. That is convenient, but it gives up control over order inside the folder. Ensure the core pieces by name first, then the category:\u003C\u002Fp>\u003Cpre>\u003Ccode># libraries and database\nensure oxmysql\nensure ox_lib\n\n# framework core\nensure qb-core\n\n# framework resources\nensure [qb]\n\n# third-party and custom\nensure [standalone]\nensure [jobs]\nensure [maps]\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>The general order for any framework is: database driver, shared libraries, framework core, inventory and targeting, then everything that depends on them. If a resource throws errors about a missing export or an undefined object on boot, the fix is almost always to move its dependency higher. A resource's \u003Ccode>fxmanifest.lua\u003C\u002Fcode> can also declare dependencies, which makes the server refuse to start it until they are running.\u003C\u002Fp>\u003Ch3>Permissions\u003C\u002Fh3>\u003Cp>Admin permissions are ACE entries, also written in the config:\u003C\u002Fp>\u003Cpre>\u003Ccode>add_ace group.admin command allow\nadd_principal identifier.fivem:1234567 group.admin\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Keep these in their own file. Framework-specific groups follow the same pattern but use the group names your framework expects, so check its documentation before copying lines from another framework's guide.\u003C\u002Fp>\u003Ch3>A layout that scales\u003C\u002Fh3>\u003Cp>Once a server passes a few dozen resources, one long file becomes hard to review. Split it:\u003C\u002Fp>\u003Col>\u003Cli>\u003Ccode>server.cfg\u003C\u002Fcode>: endpoints, identity, build, slots and the \u003Ccode>exec\u003C\u002Fcode> lines.\u003C\u002Fli>\u003Cli>\u003Ccode>secrets.cfg\u003C\u002Fcode>: license key, database string, any API keys. Never committed to Git, never pasted in Discord.\u003C\u002Fli>\u003Cli>\u003Ccode>permissions.cfg\u003C\u002Fcode>: ACE and principals.\u003C\u002Fli>\u003Cli>\u003Ccode>resources.cfg\u003C\u002Fcode>: every \u003Ccode>ensure\u003C\u002Fcode>, grouped and commented.\u003C\u002Fli>\u003C\u002Fol>\u003Cpre>\u003Ccode>exec secrets.cfg\nexec permissions.cfg\nexec resources.cfg\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Put \u003Ccode>secrets.cfg\u003C\u002Fcode> in your \u003Ccode>.gitignore\u003C\u002Fcode> if the server folder is under version control. When you ask for help, you can share the other three files without redacting anything.\u003C\u002Fp>\u003Ch3>Debugging a config that will not boot\u003C\u002Fh3>\u003Cul>\u003Cli>Read the console from the top, not the bottom. The first error is usually the real one; later errors are fallout.\u003C\u002Fli>\u003Cli>Watch for smart quotes. Copying from a website or document can turn \u003Ccode>\"\u003C\u002Fcode> into curly quotes, which the parser will not read as quotes.\u003C\u002Fli>\u003Cli>A line that does nothing is often a typo in a resource name. The console reports resources it could not find.\u003C\u002Fli>\u003Cli>Comment out half the \u003Ccode>ensure\u003C\u002Fcode> lines with \u003Ccode>#\u003C\u002Fcode> to bisect a crashing resource instead of guessing.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Takeaway\u003C\u002Fh3>\u003Cp>Treat \u003Ccode>server.cfg\u003C\u002Fcode> as a startup script that runs top to bottom. Use \u003Ccode>set\u003C\u002Fcode> for anything secret, \u003Ccode>sets\u003C\u002Fcode> only for public info, and \u003Ccode>setr\u003C\u002Fcode> only for values clients need. Start the database driver and libraries before the framework, and the framework before everything else. Split secrets, permissions and resources into separate files so the main config fits on one screen and can be shared safely. If you are setting up from scratch, the official \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\u002Fdocs\u002Fserver-manual\u002Fsetting-up-a-server\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">setting up a server guide\u003C\u002Fa> is the reference to check current requirements against.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-server-cfg-explained.webp?v=mwf7uecu",27,"FiveM server.cfg Explained Line by Line","Understand every important line in your FiveM server.cfg: endpoints, set vs sets vs setr, license key, oxmysql string, ensure order and permissions.",[19,20,21,49],"Standalone","2026-11-14T17:09:51.083+00:00","2026-11-21T07:14:36.636261+00:00",{"slug":53,"title":54,"excerpt":55,"content":56,"image":57,"category":11,"author":12,"status":13,"reviewed":14,"views":58,"metaTitle":59,"metaDescription":60,"frameworks":61,"productId":22,"publishedAt":62,"updatedAt":63,"date":62},"fivem-roleplay-economy-sources-sinks-inflation","Designing a FiveM Roleplay Economy: Sources, Sinks and Inflation","How to map where money enters and leaves your FiveM server, price jobs by time, add sinks players accept, and catch inflation before it ruins your roleplay.","\u003Cp>Most FiveM roleplay economies fail the same way. A few weeks after launch, money stops meaning anything, every second car in Legion Square is a supercar, and new players cannot catch up with anyone. That is rarely a scripting bug. It is a design problem, and you can fix it with a money map, a few targets and a habit of checking your numbers.\u003C\u002Fp>\u003Ch3>Start with a money map\u003C\u002Fh3>\u003Cp>Before you tune a single payout, write down every way money enters and leaves your server. Most owners never do this, which is why they cannot explain where inflation comes from.\u003C\u002Fp>\u003Cp>\u003Cstrong>Sources\u003C\u002Fstrong> are anything that creates money out of nothing:\u003C\u002Fp>\u003Cul>\u003Cli>Legal job payouts (delivery, garbage, bus, mining, fishing)\u003C\u002Fli>\u003Cli>Paychecks for government jobs and on-duty salaries\u003C\u002Fli>\u003Cli>Selling items to NPC shops\u003C\u002Fli>\u003Cli>Illegal activities: drug sales to NPCs, robberies, heists\u003C\u002Fli>\u003Cli>Starting cash and any admin or event rewards\u003C\u002Fli>\u003C\u002Ful>\u003Cp>\u003Cstrong>Sinks\u003C\u002Fstrong> are anything that destroys money:\u003C\u002Fp>\u003Cul>\u003Cli>Vehicle, house and item purchases from NPC shops\u003C\u002Fli>\u003Cli>Repairs, fuel, impound and towing fees\u003C\u002Fli>\u003Cli>Rent, property tax, business licenses\u003C\u002Fli>\u003Cli>Fines and hospital bills paid to the state, not to a player\u003C\u002Fli>\u003Cli>Crafting materials bought from NPCs\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Player-to-player trades are neither. Buying a car from another player moves money around but does not change the total. That distinction matters: if your only big sink is a player-owned dealership where the money goes to the owner, you have no real sink at all.\u003C\u002Fp>\u003Ch3>Price jobs by time, not by task\u003C\u002Fh3>\u003Cp>Setting a delivery to pay a round number because it feels right is how you end up with one job paying five times another. Work backwards from an hourly target instead.\u003C\u002Fp>\u003Col>\u003Cli>Decide what a new player should earn per hour of honest legal work. Pick a number relative to your prices, for example enough that a modest first car takes three to four hours of play.\u003C\u002Fli>\u003Cli>Time each job yourself, including travel, loading and the walk back to the vehicle. Do it at least three times.\u003C\u002Fli>\u003Cli>Divide the hourly target by the number of runs per hour. That is your per-run pay.\u003C\u002Fli>\u003Cli>Adjust for risk and effort. Jobs that need a group, special equipment or carry danger can pay more per hour, but within a defined band, not without limits.\u003C\u002Fli>\u003C\u002Fol>\u003Cp>Keep a simple spreadsheet with job name, minutes per run, pay per run and resulting hourly rate. When someone reports that a job is \"broken\", you can check whether it is outside its band in seconds.\u003C\u002Fp>\u003Ch4>Illegal income needs a ceiling too\u003C\u002Fh4>\u003Cp>Criminal activities should pay better per hour than legal work, because they carry risk of arrest, losing items and time spent in prison. The mistake is letting that premium grow unchecked. If selling drugs to NPCs pays ten times a legal job with little real risk, your entire population turns into dealers and police get overwhelmed. Keep illegal hourly income in a band, add cooldowns, and make sure the risk is real: active police, item loss on arrest and limited NPC demand.\u003C\u002Fp>\u003Ch3>Sinks players do not resent\u003C\u002Fh3>\u003Cp>Players hate taxes that feel like punishment. They accept costs that feel like part of the world. A few that tend to work:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Maintenance:\u003C\u002Fstrong> vehicles that slowly need repairs, fuel that costs something, tyres and parts that wear. Ownership has an ongoing cost, which is realistic.\u003C\u002Fli>\u003Cli>\u003Cstrong>Property upkeep:\u003C\u002Fstrong> weekly rent or tax on houses and businesses. This also frees inactive properties for active players if you reclaim them after missed payments.\u003C\u002Fli>\u003Cli>\u003Cstrong>Status items:\u003C\u002Fstrong> expensive clothing, unique plates, custom paint and high-end apartments. Wealthy players spend voluntarily and remove a lot of money.\u003C\u002Fli>\u003Cli>\u003Cstrong>Convenience fees:\u003C\u002Fstrong> paying for fast travel, instant impound retrieval or a private garage slot instead of waiting.\u003C\u002Fli>\u003Cli>\u003Cstrong>Licenses:\u003C\u002Fstrong> driving, weapon, business and hunting licenses that also give police something to check during roleplay.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Avoid sinks that hit poor players hardest, such as flat daily fees on everyone. Scale costs with what the player owns.\u003C\u002Fp>\u003Ch3>Inflation warning signs\u003C\u002Fh3>\u003Cp>You do not need complex analytics. Check these regularly, weekly on a new server and monthly later on:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Total money in circulation:\u003C\u002Fstrong> sum all cash and bank balances. If it grows much faster than your active player count, sources are outrunning sinks.\u003C\u002Fli>\u003Cli>\u003Cstrong>Top balances:\u003C\u002Fstrong> look at the richest 10 characters. A single account far above the rest usually means an exploit, a duplication bug or a payout that is far too generous.\u003C\u002Fli>\u003Cli>\u003Cstrong>Expensive vehicle ownership:\u003C\u002Fstrong> if the most expensive tier of cars becomes common, prices no longer mean anything.\u003C\u002Fli>\u003Cli>\u003Cstrong>Player-set prices:\u003C\u002Fstrong> when players start pricing services in huge numbers, the currency has lost value.\u003C\u002Fli>\u003Cli>\u003Cstrong>New player complaints:\u003C\u002Fstrong> \"there is no point working, everything costs too much\" often means veterans have pushed player market prices out of reach.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Log every transaction that matters\u003C\u002Fh3>\u003Cp>You cannot fix what you cannot see. Make sure every money addition from a script runs on the server and leaves a trace with the resource name and a reason. Most frameworks let you pass a reason when adding or removing money; use it consistently and send larger amounts to a log channel or database table.\u003C\u002Fp>\u003Cp>A server-side rate limit on payouts also protects you from both exploits and badly tuned jobs. A framework-agnostic sketch:\u003C\u002Fp>\u003Cpre>\u003Ccode>-- server side, example values only\nlocal HOURLY_LIMIT = 25000\nlocal earned = {}\n\nlocal function canPay(src, amount)\n    local now = os.time()\n    local entry = earned[src]\n    if not entry or now - entry.start &gt;= 3600 then\n        entry = { start = now, total = 0 }\n        earned[src] = entry\n    end\n    if entry.total + amount &gt; HOURLY_LIMIT then\n        return false\n    end\n    entry.total = entry.total + amount\n    return true\nend\n\nAddEventHandler('playerDropped', function()\n    earned[source] = nil\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>When \u003Ccode>canPay\u003C\u002Fcode> returns false, do not silently drop the payout. Log it. A legitimate player hitting the limit means your limit or your job tuning is wrong; a player hitting it repeatedly through an unusual path is worth investigating. Pair this with the basics of never trusting client-sent amounts: the server decides how much a job pays, the client only says \"I finished\".\u003C\u002Fp>\u003Ch3>Changing the economy on a live server\u003C\u002Fh3>\u003Cp>At some point you will need to cut a payout or raise a price. Done badly, this causes more drama than any other update.\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Announce first.\u003C\u002Fstrong> Give players a few days of notice and explain why. \"Fishing paid three times more per hour than any other legal job\" is easier to accept than a silent nerf.\u003C\u002Fli>\u003Cli>\u003Cstrong>Change in steps.\u003C\u002Fstrong> Two smaller adjustments a week apart are better than one big cut.\u003C\u002Fli>\u003Cli>\u003Cstrong>Raise sinks before cutting sources.\u003C\u002Fstrong> Adding a desirable new item or property soaks up money without making anyone feel poorer.\u003C\u002Fli>\u003Cli>\u003Cstrong>Test on a copy.\u003C\u002Fstrong> Run a development server with a recent database backup and check that the new values behave as expected.\u003C\u002Fli>\u003Cli>\u003Cstrong>Write it down.\u003C\u002Fstrong> Keep a changelog of economy changes with dates so you can connect them to the numbers you track.\u003C\u002Fli>\u003C\u002Ful>\u003Ch4>About wipes\u003C\u002Fh4>\u003Cp>A full economy wipe is the nuclear option. It resets inflation, but you lose players who spent months building their characters. If you must wipe, consider a partial reset: money above a cap is reduced, owned properties and vehicles are kept, and you announce it well ahead of time. Better still, design sinks early enough that a wipe is never needed.\u003C\u002Fp>\u003Ch3>Practical takeaway\u003C\u002Fh3>\u003Col>\u003Cli>List every money source and sink on your server and remove anything you cannot account for.\u003C\u002Fli>\u003Cli>Set an hourly income band for new legal work, experienced legal work and illegal work, then price every job against it.\u003C\u002Fli>\u003Cli>Add ongoing ownership costs and voluntary luxury purchases as your main sinks.\u003C\u002Fli>\u003Cli>Keep all money changes server-side, with a reason, and log the large ones.\u003C\u002Fli>\u003Cli>Check total circulation and top balances on a schedule.\u003C\u002Fli>\u003Cli>Change values gradually, with notice, and test on a copy of your live data first.\u003C\u002Fli>\u003C\u002Fol>\u003Cp>An economy is not something you configure once at launch. Treat it like any other part of server maintenance, and your money will keep meaning something to the players who earn it.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-roleplay-economy-sources-sinks-inflation.webp?v=mw6jt3j1",24,"FiveM Roleplay Economy Design: Sources and Sinks","A practical guide to FiveM server economy design: money sources and sinks, job pay by hourly targets, inflation warning signs, logging and safe live changes.",[19,20,21],"2026-11-11T20:02:01.428+00:00","2026-11-21T00:11:37.65734+00:00",{"slug":65,"title":66,"excerpt":67,"content":68,"image":69,"category":11,"author":12,"status":13,"reviewed":14,"views":70,"metaTitle":71,"metaDescription":72,"frameworks":73,"productId":22,"publishedAt":74,"updatedAt":75,"date":74},"fivem-nui-lua-messages-callbacks-focus","FiveM NUI in Lua: Messages, Callbacks and Focus Done Right","How Lua and your HTML UI actually talk in FiveM: SendNUIMessage, NUI callbacks, focus handling, security and the mistakes that leave players stuck.","\u003Cp>Most broken FiveM UIs fail in the same handful of ways. A file is missing from the manifest, a callback never answers, focus gets stuck, or the server trusts whatever the page sends it. This article walks through the NUI pipeline end to end, with a minimal pattern you can reuse in every resource.\u003C\u002Fp>\u003Ch3>The four moving parts\u003C\u002Fh3>\u003Cp>Every NUI resource is built from the same pieces, whatever framework you run:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>The manifest\u003C\u002Fstrong> declares which HTML page the resource loads and which files the page is allowed to request.\u003C\u002Fli>\u003Cli>\u003Cstrong>Lua to JavaScript\u003C\u002Fstrong>: the client script calls \u003Ccode>SendNUIMessage\u003C\u002Fcode>, and the page receives it as a standard browser \u003Ccode>message\u003C\u002Fcode> event.\u003C\u002Fli>\u003Cli>\u003Cstrong>JavaScript to Lua\u003C\u002Fstrong>: the page sends a \u003Ccode>fetch\u003C\u002Fcode> POST to \u003Ccode>https:\u002F\u002F&lt;resource-name&gt;\u002F&lt;callback&gt;\u003C\u002Fcode>, and Lua handles it with \u003Ccode>RegisterNUICallback\u003C\u002Fcode>.\u003C\u002Fli>\u003Cli>\u003Cstrong>Focus\u003C\u002Fstrong>: \u003Ccode>SetNuiFocus\u003C\u002Fcode> decides whether the page receives the mouse and keyboard.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>If you keep these four separate in your head, debugging becomes a matter of finding which link in the chain broke.\u003C\u002Fp>\u003Ch3>The manifest\u003C\u002Fh3>\u003Cpre>\u003Ccode>fx_version 'cerulean'\ngame 'gta5'\nlua54 'yes'\n\nclient_script 'client.lua'\nserver_script 'server.lua'\n\nui_page 'html\u002Findex.html'\n\nfiles {\n  'html\u002Findex.html',\n  'html\u002Fstyle.css',\n  'html\u002Fapp.js'\n}\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Every file the page loads, including fonts, images and sounds, must be covered by \u003Ccode>files\u003C\u002Fcode>. If it is not, the request fails and you usually get a blank or unstyled UI with no Lua error at all. Glob patterns such as \u003Ccode>'html\u002F**\u002F*'\u003C\u002Fcode> work, but be careful not to ship source folders, maps or \u003Ccode>node_modules\u003C\u002Fcode> by accident. If you build with Vite or similar, point the manifest at the build output only.\u003C\u002Fp>\u003Ch3>Lua to JavaScript: one message shape\u003C\u002Fh3>\u003Cp>Pick a single message format and stick with it. An \u003Ccode>action\u003C\u002Fcode> field plus a \u003Ccode>data\u003C\u002Fcode> payload keeps the JavaScript side a simple router.\u003C\u002Fp>\u003Cpre>\u003Ccode>-- client.lua\nlocal isOpen = false\n\nlocal function setOpen(state, payload)\n  isOpen = state\n  SetNuiFocus(state, state)\n  SendNUIMessage({\n    action = state and 'open' or 'close',\n    data = payload\n  })\nend\n\nRegisterCommand('shopui', function()\n  if isOpen then return end\n  setOpen(true, { title = 'General Store' })\nend, false)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cpre>\u003Ccode>\u002F\u002F app.js\nconst app = document.getElementById('app');\n\nwindow.addEventListener('message', (event) =&gt; {\n  const { action, data } = event.data;\n\n  if (action === 'open') {\n    app.classList.remove('hidden');\n    render(data);\n  } else if (action === 'close') {\n    app.classList.add('hidden');\n  }\n});\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Everything you pass to \u003Ccode>SendNUIMessage\u003C\u002Fcode> is serialized to JSON, so stick to plain tables, strings, numbers and booleans. Functions and vectors will not arrive the way you expect; convert coordinates to \u003Ccode>{ x = ..., y = ... }\u003C\u002Fcode> tables first.\u003C\u002Fp>\u003Ch3>JavaScript to Lua: callbacks that always answer\u003C\u002Fh3>\u003Cp>On the page side, wrap the request in a tiny helper. \u003Ccode>GetParentResourceName()\u003C\u002Fcode> is injected into the NUI page and returns the resource name, so you never hardcode it.\u003C\u002Fp>\u003Cpre>\u003Ccode>\u002F\u002F app.js\nasync function post(name, body = {}) {\n  const res = await fetch(`https:\u002F\u002F${GetParentResourceName()}\u002F${name}`, {\n    method: 'POST',\n    headers: { 'Content-Type': 'application\u002Fjson; charset=UTF-8' },\n    body: JSON.stringify(body)\n  });\n  return res.json();\n}\n\ndocument.getElementById('buy').addEventListener('click', () =&gt; {\n  post('buy', { item: 'water', amount: 2 });\n});\u003C\u002Fcode>\u003C\u002Fpre>\u003Cpre>\u003Ccode>-- client.lua\nRegisterNUICallback('close', function(_, cb)\n  setOpen(false)\n  cb({})\nend)\n\nRegisterNUICallback('buy', function(data, cb)\n  TriggerServerEvent('myshop:buy', data.item, data.amount)\n  cb({ ok = true })\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Two rules matter here:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Always call \u003Ccode>cb\u003C\u002Fcode>\u003C\u002Fstrong>, on every code path, including early returns. If you forget, the \u003Ccode>fetch\u003C\u002Fcode> promise hangs and any UI logic waiting on it freezes.\u003C\u002Fli>\u003Cli>\u003Cstrong>Answer with a table\u003C\u002Fstrong>, even an empty one. If the page calls \u003Ccode>res.json()\u003C\u002Fcode> and you replied with a bare string like \u003Ccode>'ok'\u003C\u002Fcode>, parsing throws and the error is easy to miss.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Focus: open cleanly, close every time\u003C\u002Fh3>\u003Cp>A stuck cursor is the bug players report most, and it is almost always a missing close path. Give the player several ways out and make all of them go through the same function.\u003C\u002Fp>\u003Cpre>\u003Ccode>\u002F\u002F app.js\ndocument.addEventListener('keyup', (e) =&gt; {\n  if (e.key === 'Escape') post('close');\n});\u003C\u002Fcode>\u003C\u002Fpre>\u003Cpre>\u003Ccode>-- client.lua\nAddEventHandler('onResourceStop', function(resource)\n  if resource == GetCurrentResourceName() and isOpen then\n    SetNuiFocus(false, false)\n  end\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>The resource stop handler matters during development: restart a resource while its UI is open and, without it, you may be left with no mouse control until you reconnect. Also close the UI when the player dies, enters a vehicle or gets cuffed, if that suits your design. A single \u003Ccode>setOpen(false)\u003C\u002Fcode> keeps state and focus consistent.\u003C\u002Fp>\u003Cp>If players should keep walking while the UI is open, look at \u003Ccode>SetNuiFocusKeepInput\u003C\u002Fcode>. Be aware that game controls then fire while the player types, so you will need to disable the relevant controls in a loop that only runs while the UI is open.\u003C\u002Fp>\u003Ch3>The page is client code, so treat it that way\u003C\u002Fh3>\u003Cp>Anything running in NUI is under the player's control. A cheater can call your callbacks with any payload they like, skip your client-side checks entirely and fire your server event directly. The page decides what to show; the server decides what happens.\u003C\u002Fp>\u003Cpre>\u003Ccode>-- server.lua\nlocal ITEMS = { water = 5, bread = 8 }\n\nRegisterNetEvent('myshop:buy', function(item, amount)\n  local src = source\n  local price = ITEMS[item]\n  amount = math.floor(tonumber(amount) or 0)\n\n  if not price or amount &lt; 1 or amount &gt; 10 then return end\n\n  -- check distance to the shop, check money through your\n  -- framework, remove money, then give the item\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cul>\u003Cli>Never send prices, rewards or permissions from the page. Look them up on the server.\u003C\u002Fli>\u003Cli>Clamp and type-check every number. \u003Ccode>tonumber\u003C\u002Fcode> and \u003Ccode>math.floor\u003C\u002Fcode> stop strings, decimals and negatives.\u003C\u002Fli>\u003Cli>Check that the player is actually where the action makes sense, using server-side ped coordinates.\u003C\u002Fli>\u003Cli>Rate-limit expensive or valuable actions per \u003Ccode>source\u003C\u002Fcode>.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Keeping NUI cheap\u003C\u002Fh3>\u003Cp>\u003Ccode>resmon\u003C\u002Fcode> shows Lua time, not browser time, so a heavy UI can drag client FPS while your resource looks clean. Watch for these:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Do not stream data every frame.\u003C\u002Fstrong> Calling \u003Ccode>SendNUIMessage\u003C\u002Fcode> inside a \u003Ccode>Wait(0)\u003C\u002Fcode> loop is a common HUD mistake. Send on change, or throttle to a few times per second.\u003C\u002Fli>\u003Cli>\u003Cstrong>Stop work when hidden.\u003C\u002Fstrong> Clear \u003Ccode>setInterval\u003C\u002Fcode> timers and pause CSS animations when the UI closes. A hidden element with a running animation still costs.\u003C\u002Fli>\u003Cli>\u003Cstrong>Compress assets.\u003C\u002Fstrong> Large PNGs and unoptimized fonts slow first open and every client download.\u003C\u002Fli>\u003Cli>\u003Cstrong>Ship a build, not a dev setup.\u003C\u002Fstrong> Minified bundles load faster and leak less of your source.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Debugging checklist\u003C\u002Fh3>\u003Cp>FiveM includes Chromium DevTools for NUI, which you can open from the F8 console with \u003Ccode>nui_devtools\u003C\u002Fcode>. The Console and Network tabs answer most questions. When something does not work, go down this list:\u003C\u002Fp>\u003Col>\u003Cli>\u003Cstrong>Blank page or missing styles\u003C\u002Fstrong>: check the Network tab for failed requests and confirm the file is listed in \u003Ccode>files\u003C\u002Fcode>.\u003C\u002Fli>\u003Cli>\u003Cstrong>Message never arrives\u003C\u002Fstrong>: log \u003Ccode>event.data\u003C\u002Fcode> in the listener and confirm the \u003Ccode>action\u003C\u002Fcode> string matches exactly.\u003C\u002Fli>\u003Cli>\u003Cstrong>Callback hangs\u003C\u002Fstrong>: find the code path in Lua that returns without calling \u003Ccode>cb\u003C\u002Fcode>.\u003C\u002Fli>\u003Cli>\u003Cstrong>Callback 404s\u003C\u002Fstrong>: the name in \u003Ccode>fetch\u003C\u002Fcode> and in \u003Ccode>RegisterNUICallback\u003C\u002Fcode> differ, or the resource folder was renamed and something is hardcoded.\u003C\u002Fli>\u003Cli>\u003Cstrong>Old version still showing\u003C\u002Fstrong>: restart the resource; if it persists, clear the client NUI cache.\u003C\u002Fli>\u003Cli>\u003Cstrong>Cursor stuck\u003C\u002Fstrong>: some close path skips \u003Ccode>SetNuiFocus(false, false)\u003C\u002Fcode>.\u003C\u002Fli>\u003C\u002Fol>\u003Cp>For native signatures, the \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\u002Fnatives\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">FiveM natives reference\u003C\u002Fa> is the source of truth.\u003C\u002Fp>\u003Ch3>Takeaway\u003C\u002Fh3>\u003Cul>\u003Cli>One message shape (\u003Ccode>action\u003C\u002Fcode> + \u003Ccode>data\u003C\u002Fcode>) and one \u003Ccode>post()\u003C\u002Fcode> helper per resource.\u003C\u002Fli>\u003Cli>Every callback calls \u003Ccode>cb\u003C\u002Fcode> with a table, on every path.\u003C\u002Fli>\u003Cli>All close paths go through one function that releases focus, plus an \u003Ccode>onResourceStop\u003C\u002Fcode> safety net.\u003C\u002Fli>\u003Cli>The page shows; the server decides. Validate type, range, distance and price server-side.\u003C\u002Fli>\u003Cli>Send NUI messages on change, stop timers when hidden and check DevTools before blaming Lua.\u003C\u002Fli>\u003C\u002Ful>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-nui-lua-messages-callbacks-focus.webp?v=mwax5kz7",19,"FiveM NUI in Lua: Messages, Callbacks and Focus","A practical guide to FiveM NUI: fxmanifest setup, SendNUIMessage, RegisterNUICallback, focus handling, security and debugging stuck or broken UIs.",[49,19,20,21],"2026-11-11T20:01:42.259+00:00","2026-11-17T20:45:51.012342+00:00",{"slug":77,"title":78,"excerpt":79,"content":80,"image":81,"category":82,"author":12,"status":13,"reviewed":83,"views":70,"metaTitle":84,"metaDescription":85,"frameworks":86,"productId":88,"publishedAt":89,"updatedAt":90,"date":89},"best-fivem-emote-menu","Best FiveM Emote Menu in 2026: What to Look For Before You Install One","Every roleplay server needs an emote menu, and players use it more than almost any other UI. Here is what separates a good one from a list of animations.","\u003Cp>The emote menu is one of the few scripts every single player on a roleplay server touches, often dozens of times per session. Leaning on a wall during a scene, sitting at a bar, handing over a coffee, dancing at a club: all of it runs through the emote menu. That makes it a surprisingly important choice. A clumsy one slows down every scene; a good one disappears into the roleplay.\u003C\u002Fp>\n\u003Cp>This guide covers what to look for when you pick a FiveM emote menu in 2026, the free options most servers start with, and when a premium menu is worth it.\u003C\u002Fp>\n\n\u003Ch2>What a FiveM emote menu actually has to do\u003C\u002Fh2>\n\u003Cp>At its core an emote menu maps a name to a GTA V animation dictionary and clip, plays it on the player's ped and stops it cleanly. Everything beyond that is about speed and expression. When you compare menus, these are the questions that matter in practice.\u003C\u002Fp>\n\n\u003Ch3>1. Can players find the right emote quickly?\u003C\u002Fh3>\n\u003Cp>A server easily ends up with several hundred emotes. Scrolling a long list mid-scene breaks immersion, so look for \u003Cstrong>search\u003C\u002Fstrong>, clear \u003Cstrong>categories\u003C\u002Fstrong> and a \u003Cstrong>favorites\u003C\u002Fstrong> list so regular players can keep their go-to animations one click away.\u003C\u002Fp>\n\n\u003Ch3>2. Can players see an emote before they play it?\u003C\u002Fh3>\n\u003Cp>Emote names like \"lean4\" or \"sitchair2\" mean nothing until you try them. A \u003Cstrong>preview\u003C\u002Fstrong> lets players check the animation before committing to it in front of other people. It sounds small, but it is the single feature players mention most once they have used it.\u003C\u002Fp>\n\n\u003Ch3>3. Does it support shared (synced) emotes?\u003C\u002Fh3>\n\u003Cp>Handshakes, hugs, high fives and partner dances need two players and a synced start. A menu with \u003Cstrong>shared emotes\u003C\u002Fstrong> sends a request to the nearby player and plays both sides together, which is what makes these interactions feel natural instead of awkward.\u003C\u002Fp>\n\n\u003Ch3>4. Can you add your own animations?\u003C\u002Fh3>\n\u003Cp>Most servers eventually stream custom animations: gang signs, job-specific poses, props for a café. Check that the menu accepts \u003Cstrong>custom emotes\u003C\u002Fstrong> through config rather than requiring code edits, so adding one is a few lines and a restart.\u003C\u002Fp>\n\n\u003Ch3>5. Does it fit your framework and your UI?\u003C\u002Fh3>\n\u003Cp>An emote menu has no real reason to depend on a framework. A \u003Cstrong>standalone\u003C\u002Fstrong> menu works the same on QBCore, ESX and Qbox and survives framework migrations. It should also look like it belongs on your server: theme options matter more here than on admin-only tools, because every player sees it.\u003C\u002Fp>\n\n\u003Ch2>The free options most servers start with\u003C\u002Fh2>\n\u003Cp>Many servers begin with a free, community-maintained emote menu such as dpemotes or one of its forks like rpemotes. They are a sensible starting point: open source, a large built-in animation list, and widely known by players coming from other servers. The trade-off is usually the interface. Classic list-style menus work, but finding a specific emote takes longer, there is often no visual preview, and restyling them to match a modern NUI takes real front-end work.\u003C\u002Fp>\n\u003Cp>If your server is small or you mostly care about the animation list itself, a free menu is fine. If players use emotes constantly and you care how the experience feels, the interface is where the upgrade pays off.\u003C\u002Fp>\n\n\u003Ch2>Checklist: choosing your emote menu\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Search, categories and favorites\u003C\u002Fli>\n\u003Cli>Preview before playing\u003C\u002Fli>\n\u003Cli>Shared emotes with a request\u002Faccept flow\u003C\u002Fli>\n\u003Cli>Custom emotes added through config\u003C\u002Fli>\n\u003Cli>Standalone, so it runs on QBCore, ESX and Qbox alike\u003C\u002Fli>\n\u003Cli>More than one UI theme, or a UI you can restyle\u003C\u002Fli>\n\u003Cli>No measurable idle cost when the menu is closed\u003C\u002Fli>\n\u003C\u002Ful>\n\n\u003Ch2>Where the ATY Emote Menu fits\u003C\u002Fh2>\n\u003Cp>We built the \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Fproduct\u002F6832535\">ATY Emote Menu\u003C\u002Fa> around exactly this checklist. Players can \u003Cstrong>preview any emote\u003C\u002Fstrong> with one click before using it, save emotes to \u003Cstrong>favorites\u003C\u002Fstrong>, and find what they need through a fast search. \u003Cstrong>Shared emotes\u003C\u002Fstrong> handle synced interactions like handshakes, hugs and dances. Server owners get \u003Cstrong>three UI themes\u003C\u002Fstrong>, configurable categories and options, and full support for \u003Cstrong>custom emotes\u003C\u002Fstrong>.\u003C\u002Fp>\n\u003Cp>It is \u003Cstrong>standalone\u003C\u002Fstrong>, so it does not care which framework you run, and it is available as an escrow resource or with full source code. If you are unsure whether it fits your server, ask us on Discord before you buy: if it doesn't run on your server as described, we fix it within 48 hours or refund you.\u003C\u002Fp>\n\n\u003Ch2>Frequently asked questions\u003C\u002Fh2>\n\u003Ch3>Do I need to remove my old emote menu first?\u003C\u002Fh3>\n\u003Cp>Yes. Run one emote menu at a time; two resources listening for the same command or keybind will conflict. Stop the old resource, start the new one and move any custom animations into its config.\u003C\u002Fp>\n\u003Ch3>Will an emote menu affect server performance?\u003C\u002Fh3>\n\u003Cp>A well-built menu does nothing while closed. Animations are played client-side by GTA itself; the menu only loads the dictionary and starts the clip, so the ongoing cost should be zero.\u003C\u002Fp>\n\u003Ch3>Does the emote menu work on Qbox?\u003C\u002Fh3>\n\u003Cp>A standalone emote menu works on Qbox, QBCore, ESX and custom frameworks, because it does not call framework functions at all.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002Fb932c9359b9a50213ba3dbd758ca8be0e13e29e6.png","Guides",true,"Best FiveM Emote Menu (2026): Features That Matter","How to choose a FiveM emote menu in 2026: previews, favorites, shared emotes, custom animations and framework support, compared for QBCore, ESX and Qbox.",[19,20,87,49],"QBox",6832535,"2026-11-09T18:29:14.071+00:00","2026-11-20T09:54:49.756809+00:00",{"slug":92,"title":93,"excerpt":94,"content":95,"image":96,"category":82,"author":12,"status":13,"reviewed":83,"views":97,"metaTitle":98,"metaDescription":99,"frameworks":100,"productId":101,"publishedAt":89,"updatedAt":102,"date":89},"ox-inventory-redesign-guide","ox_inventory Redesign: How to Upgrade the ox_inventory UI (and When to Replace It)","ox_inventory is the default inventory for a huge share of FiveM servers. Its logic is solid, its default look is not for everyone. Here is how to approach a redesign.","\u003Cp>ox_inventory, by the Overextended team, has become the default inventory for a large share of FiveM servers. It is free, open source, slot based, and works with ESX, QBCore and Qbox. Its server-side item logic is solid, which is exactly why so many server owners keep it. What they often want to change is how it looks: the default interface is functional, but it is the same interface players have seen on dozens of other servers.\u003C\u002Fp>\n\u003Cp>That is what an ox_inventory redesign is for. This guide explains what a redesign can change, what it should leave alone, and when you would be better served by replacing the inventory entirely.\u003C\u002Fp>\n\n\u003Ch2>What an ox_inventory redesign changes\u003C\u002Fh2>\n\u003Cp>A redesign replaces the \u003Cstrong>web UI\u003C\u002Fstrong> (the NUI) that ox_inventory renders, while ox_inventory itself keeps handling items, stashes, shops, drops and crafting on the server. Done properly, that means:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Your existing items, stashes and player inventories stay exactly where they are\u003C\u002Fli>\n\u003Cli>Scripts that call ox_inventory exports keep working, because the exports do not change\u003C\u002Fli>\n\u003Cli>You can switch back to the original UI if you ever need to\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>What a redesign \u003Cstrong>cannot\u003C\u002Fstrong> change is the underlying item model. If you want behavior ox_inventory does not have, a new skin will not add it.\u003C\u002Fp>\n\n\u003Ch2>What to check before installing a redesign\u003C\u002Fh2>\n\u003Ch3>Does it keep ox_inventory up to date?\u003C\u002Fh3>\n\u003Cp>ox_inventory receives updates. A redesign that forks and edits the whole resource can leave you stuck on an old version. Prefer one that changes the UI layer and documents exactly which files it touches.\u003C\u002Fp>\n\u003Ch3>Does it add things players actually notice?\u003C\u002Fh3>\n\u003Cp>The changes that make a redesign feel worth it are visible ones: a live character preview inside the inventory, clothing slots, readable status bars, clear rarity styling and crisp text over a blurred background.\u003C\u002Fp>\n\u003Ch3>Does it handle backpacks properly?\u003C\u002Fh3>\n\u003Cp>Backpacks are the most requested addition. The details matter: a backpack's contents must persist when it is dropped or given away, and you need to block bag-in-bag nesting or players will carry unlimited storage.\u003C\u002Fp>\n\u003Ch3>What does it cost at runtime?\u003C\u002Fh3>\n\u003Cp>Visual effects like background blur can be expensive if done carelessly. Open the inventory with resmon running and check the client cost while it is open and closed.\u003C\u002Fp>\n\n\u003Ch2>The ATY ox_inventory Redesign\u003C\u002Fh2>\n\u003Cp>Our \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Fproduct\u002F7357848\">ox_inventory Redesign\u003C\u002Fa> follows that checklist. It adds a \u003Cstrong>live 3D character preview\u003C\u002Fstrong> in the center of the inventory, an \u003Cstrong>integrated clothing system\u003C\u002Fstrong> for managing outfits from inventory slots, and redesigned \u003Cstrong>hunger and thirst bars\u003C\u002Fstrong>. Its \u003Cstrong>backpack system\u003C\u002Fstrong> ships in three sizes with isolated, persistent stashes: items stay saved even after the bag is dropped or given away, and \u003Cstrong>nesting prevention\u003C\u002Fstrong> stops bag-in-bag storage.\u003C\u002Fp>\n\u003Cp>The background uses a WebGL blur, with icons and text kept sharp on top of it, and item tiers get distinct rarity borders and glow. It ships as open source, so you can adjust anything you like.\u003C\u002Fp>\n\n\u003Ch2>When to replace the inventory instead\u003C\u002Fh2>\n\u003Cp>A redesign is the right call when you are happy with how ox_inventory behaves. Consider replacing it when you need behavior it was not built around, for example:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Weapons with serial numbers, ownership history and per-shot durability\u003C\u002Fli>\n\u003Cli>Food that spoils on a real clock and tools that wear down and break\u003C\u002Fli>\n\u003Cli>Items that cannot be dropped or traded, like ID cards and licences\u003C\u002Fli>\n\u003Cli>A ledger of every item movement, so you can trace and roll back duplication\u003C\u002Fli>\n\u003Cli>An in-game admin panel for items, containers and players\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>That is the problem our full \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Finventory\">ATY Inventory\u003C\u002Fa> solves. It is server-authoritative, re-validating every drag, split, trade and craft on the server, writes every change to a ledger, and ships a migration tool that dry-runs an import from ox_inventory or qb-inventory before writing anything. Scripts that call ox_inventory or qb-inventory exports keep working through a compatibility layer.\u003C\u002Fp>\n\n\u003Ch2>Frequently asked questions\u003C\u002Fh2>\n\u003Ch3>Will a redesign wipe my players' items?\u003C\u002Fh3>\n\u003Cp>No. A UI redesign only changes what players see. Items live in ox_inventory's own data, which the redesign does not touch.\u003C\u002Fp>\n\u003Ch3>Does the redesign work on QBCore, ESX and Qbox?\u003C\u002Fh3>\n\u003Cp>It works wherever ox_inventory works, and ox_inventory supports all three.\u003C\u002Fp>\n\u003Ch3>Can I use the redesign and the ATY Inventory together?\u003C\u002Fh3>\n\u003Cp>No, they solve the same problem in two different ways. Use the redesign to keep ox_inventory, or the ATY Inventory to replace it.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002Fead417c7329806ad1cc8aeeedd27a142a18c1264.jpg",20,"ox_inventory Redesign Guide (2026): UI Upgrade vs Replacement","Thinking about an ox_inventory redesign? What a new UI can and cannot change, what to check before installing one, and when a full FiveM inventory replacement makes more sense.",[19,20,87],7357848,"2026-11-20T13:26:06.551254+00:00",{"slug":104,"title":105,"excerpt":106,"content":107,"image":108,"category":82,"author":12,"status":13,"reviewed":83,"views":58,"metaTitle":109,"metaDescription":110,"frameworks":111,"productId":112,"publishedAt":89,"updatedAt":113,"date":89},"fivem-map-editor-prop-placer-guide","FiveM Map Editor & Prop Placer: Building Maps In Game vs CodeWalker (2026)","Most FiveM mapping still happens in CodeWalker on a desktop. In-game map editors and prop placers change that workflow. Here is when each one makes sense.","\u003Cp>For years, FiveM mapping meant one workflow: open CodeWalker on your desktop, place props in an offline copy of the world, export a YMAP, upload it to the server, restart, join and check whether it looks right. It is powerful, and for large MLO work it is still the right tool. But for the everyday job of decorating a car meet, adding barriers to an event or building a gang hideout, it is slow.\u003C\u002Fp>\n\u003Cp>In-game map editors, often called prop placers, move that work onto the live server. This guide explains how they compare with CodeWalker and what to look for when you pick one.\u003C\u002Fp>\n\n\u003Ch2>CodeWalker vs an in-game map editor\u003C\u002Fh2>\n\u003Ch3>CodeWalker\u003C\u002Fh3>\n\u003Cp>CodeWalker is a free desktop tool that reads the game files directly. It is the standard for editing YMAPs, collisions and interiors, and nothing in-game replaces it for deep map work. The cost is the loop: every change means export, upload, restart and rejoin, and you never see your build with real players, lighting and traffic until afterwards.\u003C\u002Fp>\n\u003Ch3>In-game map editor\u003C\u002Fh3>\n\u003Cp>An in-game editor runs as a resource on your server. You fly a camera through the live world, place and adjust props, and see the result exactly as players will. Changes can be saved without a restart, and staff can build together. The trade-off is scope: an in-game editor places and manages props and lights; it does not edit interiors or collisions the way CodeWalker does.\u003C\u002Fp>\n\u003Cp>Most servers end up using both: CodeWalker for structural map work, an in-game editor for everything built on top of the world.\u003C\u002Fp>\n\n\u003Ch2>What to look for in a FiveM prop placer\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>A real gizmo\u003C\u002Fstrong> for move, rotate and scale, in both world and local axes, with ground alignment\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Good prop search\u003C\u002Fstrong>, because GTA V has thousands of models with unhelpful names, plus favorites and your own categories\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Bulk tools\u003C\u002Fstrong> to fill an area with props or clear one in a single action\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Removing default GTA objects\u003C\u002Fstrong> that block your build, permanently and for everyone\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Lights\u003C\u002Fstrong> you can place and tune, not just props\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Persistence\u003C\u002Fstrong>: builds survive restarts without manual steps\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Export\u003C\u002Fstrong> to formats other tools understand, so your work is never locked in\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Permissions\u003C\u002Fstrong> tied to your framework, so only staff can open it\u003C\u002Fli>\n\u003C\u002Ful>\n\n\u003Ch2>The ATY Map Editor\u003C\u002Fh2>\n\u003Cp>The \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Fmap-editor\">ATY Map Editor &amp; Prop Placer\u003C\u002Fa> covers that list. It has an \u003Cstrong>advanced gizmo\u003C\u002Fstrong> for translation, rotation and scaling in world and local space, 3D raycast placement with \u003Cstrong>ground alignment\u003C\u002Fstrong>, and a smart prop search with custom categories and favorites. You can swap the model of a placed prop without re-placing it.\u003C\u002Fp>\n\u003Cp>For bulk work there is \u003Cstrong>area fill\u003C\u002Fstrong> (grid or random distribution) and \u003Cstrong>area delete\u003C\u002Fstrong>, and the \u003Cstrong>world eraser\u003C\u002Fstrong> permanently suppresses default GTA V objects, with every removal listed so it can be restored. A \u003Cstrong>light editor\u003C\u002Fstrong> handles spot and point lights with shadows, intensity and color, and an activity log records what was changed.\u003C\u002Fp>\n\u003Cp>Builds are stored on the server and rebuild themselves after a restart, with no database to set up. When you need your work elsewhere, export it as a \u003Cstrong>CodeWalker XML\u002FYMAP\u003C\u002Fstrong>, a \u003Cstrong>standalone Lua resource\u003C\u002Fstrong>, a \u003Cstrong>Sollumz Blender\u003C\u002Fstrong> script or plain JSON. Access runs through ACE, QBCore permissions or ESX groups.\u003C\u002Fp>\n\n\u003Ch2>Frequently asked questions\u003C\u002Fh2>\n\u003Ch3>Do players need the editor running to see my map?\u003C\u002Fh3>\n\u003Cp>No. Export the build as a resource, ensure it in server.cfg and it streams like any other map. The editor itself can be stopped.\u003C\u002Fp>\n\u003Ch3>Can an in-game editor replace CodeWalker completely?\u003C\u002Fh3>\n\u003Cp>Not for interiors, collisions or MLO work. Use it for everything you build on top of the world, and keep CodeWalker for structural changes.\u003C\u002Fp>\n\u003Ch3>Can I remove props that come from another map resource?\u003C\u002Fh3>\n\u003Cp>Yes. The world eraser and mover work on props from GTA itself and from other map resources you run. Building shells and terrain are map geometry, not props, so those cannot be picked.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002F7ce109982ed378ed26ecb891060f03f78ff219f1.png","FiveM Map Editor & Prop Placer Guide (2026)","In-game FiveM map editor or CodeWalker? How prop placers work, what to look for in 2026, and how to export your build as a YMAP or a standalone resource.",[19,20,87,49],7471609,"2026-11-20T10:16:44.449581+00:00",{"slug":115,"title":116,"excerpt":117,"content":118,"image":119,"category":31,"author":12,"status":13,"reviewed":14,"views":120,"metaTitle":121,"metaDescription":122,"frameworks":123,"productId":124,"publishedAt":125,"updatedAt":126,"date":125},"fivem-advanced-mechanic-tuning-work-orders","FiveM Advanced Mechanic: From Live Tuning Preview to Work Orders","How ATY Mechanic turns tuning into a real job: live previews for customers, quotes and work orders for mechanics, plus wear, towing, plates and an underground shop.","\u003Cp>On most roleplay servers the mechanic job is a tuning menu with a job check in front of it. A player drives in, someone opens the menu, clicks a few upgrades and the car leaves. There is nothing to do between the click and the payment, so mechanics get bored and customers skip them entirely. \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Fproduct\u002F7775407\">ATY Advanced Mechanic\u003C\u002Fa> is built to fix that by turning tuning into an actual workflow, from the customer's first look at a part to the mechanic bolting it on.\u003C\u002Fp>\u003Ch3>The problem with menu-only mechanic jobs\u003C\u002Fh3>\u003Cp>If the menu does all the work, the job has no gameplay. That shows up in a few predictable ways:\u003C\u002Fp>\u003Cul>\u003Cli>Mechanics are only needed because the menu is locked behind a job, not because they add anything.\u003C\u002Fli>\u003Cli>Cars never break in interesting ways, so nobody comes back after the first visit.\u003C\u002Fli>\u003Cli>Tow trucks, plates and police vehicle checks live in separate scripts that do not talk to each other.\u003C\u002Fli>\u003Cli>Prices and payouts are often calculated on the client, which invites abuse.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>ATY Mechanic bundles the tuning menu, the job, wear, diagnostics, towing and plate checks into one resource, and decides prices, payouts, stock and fill levels on the server.\u003C\u002Fp>\u003Ch3>What the customer sees: the cinematic tuning menu\u003C\u002Fh3>\u003Cp>The customer side starts with a home screen that lists every part, with search and a live preview. The camera frames each part as you browse it, so changing the exhaust shows the exhaust, not a wide shot of the whole car. It works with mouse and keyboard: hover, click, scroll, switch sections with Q and E, and hold to buy so nobody spends money with a stray click.\u003C\u002Fp>\u003Cp>What is in the menu:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Cosmetics:\u003C\u002Fstrong> 69 paints in 9 families, custom colour, pearl, neon, xenon, tyre smoke, liveries, horns and extras.\u003C\u002Fli>\u003Cli>\u003Cstrong>Wheels and stance:\u003C\u002Fstrong> custom and drift tyres, rim fit, camber, track and lowering sliders.\u003C\u002Fli>\u003Cli>\u003Cstrong>Custom plates:\u003C\u002Fstrong> changing the plate text moves the registration, garage entry, keys and service book with it, so a new plate does not orphan the car.\u003C\u002Fli>\u003Cli>\u003Cstrong>Performance:\u003C\u002Fstrong> engine, brakes, gearbox, suspension, armour, turbo, air intake, engine sounds, anti-lag and driving profiles.\u003C\u002Fli>\u003Cli>\u003Cstrong>ECU map:\u003C\u002Fstrong> boost, timing and mixture sliders that trade power gain against engine stress, with knock and blown engines if you push too far.\u003C\u002Fli>\u003Cli>\u003Cstrong>Extras:\u003C\u002Fstrong> nitro, launch control, air suspension and a manual sequential gearbox.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>There is also a built-in dyno with power and torque curves, before and after comparisons, peaks and power band. That gives tuners something to argue about and a reason to come back for another map.\u003C\u002Fp>\u003Ch3>How a job flows from quote to lift\u003C\u002Fh3>\u003Cp>The important part is that the menu does not have to be the end of the interaction. When mechanics are on duty, a typical visit looks like this:\u003C\u002Fp>\u003Col>\u003Cli>The customer builds a cart in the tuning menu, seeing every change live on their car.\u003C\u002Fli>\u003Cli>A mechanic quotes the cart, either in person or from the mechanic tablet.\u003C\u002Fli>\u003Cli>The accepted quote becomes a real work order.\u003C\u002Fli>\u003Cli>The mechanic takes the parts from the shelf, puts the car on the two-post lift or a jack, and installs parts one by one through install minigames.\u003C\u002Fli>\u003Cli>Several mechanics can work on the same car, which makes busy shops feel like an actual garage.\u003C\u002Fli>\u003C\u002Fol>\u003Cp>When nobody is on duty, an NPC mechanic covers the shop so players are not stuck waiting. Repairs are handled per part, including window glass, and repair and tyre kits exist for fixes in the field.\u003C\u002Fp>\u003Ch3>Wear, diagnostics and the reason to come back\u003C\u002Fh3>\u003Cp>A mechanic job only has repeat business if cars degrade. ATY Mechanic tracks oil, plugs, coolant, battery, brake pads, clutch and tyres, and players feel it while driving: misfires, overheating and weak brakes. Neglect a car and it drives like it.\u003C\u002Fp>\u003Cp>Mechanics get tools to work with that:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>OBD scanner:\u003C\u002Fstrong> fault codes, wear, installed performance parts and the service book.\u003C\u002Fli>\u003Cli>\u003Cstrong>Inspection reports\u003C\u002Fstrong> for documenting a car's state.\u003C\u002Fli>\u003Cli>\u003Cstrong>Vehicle card:\u003C\u002Fstrong> stats, condition, wear and service history in one place.\u003C\u002Fli>\u003Cli>\u003Cstrong>Odometer tampering\u003C\u002Fstrong> that a mechanic's scan can catch, which opens up used-car scams and the investigations that follow.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Stock is part of the loop too. Shops hold parts, restock through supply runs with a delivery van, and can produce items at a crafting bench.\u003C\u002Fp>\u003Ch3>Towing, plates and police\u003C\u002Fh3>\u003Cp>Towing is included rather than bolted on. Mechanics get a flatbed with ramp, winch and straps, hook trucks and cones. Players can call for a tow with \u003Ccode>\u002Fcalltow\u003C\u002Fcode>, there are NPC tow jobs for quiet hours, and impound is handled through the garage integration.\u003C\u002Fp>\u003Cp>On the law enforcement side, plates and vehicles become evidence:\u003C\u002Fp>\u003Cul>\u003Cli>Plates can be taken off and fake plates fitted.\u003C\u002Fli>\u003Cli>VINs can be checked, and ground off by criminals.\u003C\u002Fli>\u003Cli>Officers can run plate queries.\u003C\u002Fli>\u003Cli>ANPR patrol readers and fixed cameras flag vehicles.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>That gives police and criminals a shared system to play against each other, with mechanics in the middle.\u003C\u002Fp>\u003Ch3>The underground workshop\u003C\u002Fh3>\u003Cp>Every server has players who want the illegal version. The underground workshop has no blip, takes dirty money only and keeps no records. It sells race nitro and an unlocked ECU. Combined with fake plates and VIN grinding, it gives criminal groups a place to build getaway cars without touching the legal shop's paperwork.\u003C\u002Fp>\u003Ch3>The staff side: tablet, grades and editor\u003C\u002Fh3>\u003Cp>Mechanics manage the business from a tablet that covers work orders, supply orders, stock, quotes, invoices, team and tow calls. Owners set grade permissions per workshop, so an apprentice at one shop does not automatically get the same access as a manager at another.\u003C\u002Fp>\u003Cp>For server staff, the in-game workshop editor lets you place every point, bay and lift without a restart. That matters if you use custom MLOs: you walk the interior and place things where they fit instead of copying coordinates into a config and restarting repeatedly.\u003C\u002Fp>\u003Cp>There are two menu themes and two tablet looks, English and Turkish locales, and an editable config, locales and open server hooks. The core is escrowed, so customisation goes through the config, locales and hooks.\u003C\u002Fp>\u003Ch3>Fitting it into your server\u003C\u002Fh3>\u003Cp>The resource runs on QBCore or ESX through aty_lib, which provides the framework bridge, callbacks and shared helpers. Requirements:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ccode>aty_lib\u003C\u002Fcode>\u003C\u002Fli>\u003Cli>\u003Ccode>aty_garage_v2\u003C\u002Fcode> 1.11.0 or newer, for garage, impound and plate changes\u003C\u002Fli>\u003Cli>A target script: \u003Ccode>ox_target\u003C\u002Fcode>, \u003Ccode>qb-target\u003C\u002Fcode> or \u003Ccode>aty_interactions\u003C\u002Fcode>\u003C\u002Fli>\u003Cli>A SQL driver: \u003Ccode>oxmysql\u003C\u002Fcode>, \u003Ccode>ghmattimysql\u003C\u002Fcode> or \u003Ccode>mysql-async\u003C\u002Fcode>\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Start dependencies before the mechanic resource. A typical order in \u003Ccode>server.cfg\u003C\u002Fcode>:\u003C\u002Fp>\u003Cpre>\u003Ccode>ensure oxmysql\nensure qb-core        # or es_extended\nensure ox_target      # or your target script\nensure aty_lib\nensure aty_garage_v2\n# then the mechanic resource\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Before opening it to players, run through this checklist:\u003C\u002Fp>\u003Col>\u003Cli>Confirm your garage resource meets the version requirement, or plate changes and impound will not line up.\u003C\u002Fli>\u003Cli>Remove or disable any existing tuning menu and tow script so players do not have two competing systems.\u003C\u002Fli>\u003Cli>Use the workshop editor to place bays, lifts and shelves in each shop you plan to run.\u003C\u002Fli>\u003Cli>Set grade permissions per workshop before hiring anyone.\u003C\u002Fli>\u003Cli>Decide where the underground workshop goes and keep it off any public map you share.\u003C\u002Fli>\u003Cli>Brief your police department on plate queries, VIN checks and ANPR so the system gets used.\u003C\u002Fli>\u003C\u002Fol>\u003Cp>If you are unsure how ensure order works, the \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\" target=\"_blank\" rel=\"noopener noreferrer\">FiveM documentation\u003C\u002Fa> covers server configuration.\u003C\u002Fp>\u003Ch3>Who it is for\u003C\u002Fh3>\u003Cp>This is a good fit for servers that want mechanics to be a real job with shifts, stock and teamwork, and for servers where car culture matters: tuners, racers and the criminals who need fast cars with clean-looking plates. It also suits servers that want police vehicle checks tied into the same data as the garage and the mechanic shop.\u003C\u002Fp>\u003Cp>If your server has no active mechanic players, the NPC mechanic and NPC tow jobs keep the systems usable, but most of the value comes from the player-driven quote and work order loop.\u003C\u002Fp>\u003Ch3>Takeaway\u003C\u002Fh3>\u003Cp>A mechanic job works when players need each other: customers need a quote and an install, mechanics need parts and stock, police need plate data, and criminals need somewhere off the books. ATY Mechanic covers all of those in one resource. Check the requirements against your server, plan your workshop layouts, and look at the full feature list on the \u003Ca href=\"https:\u002F\u002Fwww.atyscripts.com\u002Fproduct\u002F7775407\">product page\u003C\u002Fa>.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-advanced-mechanic-tuning-work-orders.webp?v=mw2ci5qd",22,"FiveM Advanced Mechanic: Tuning Menu & Mechanic Job","A walkthrough of ATY Mechanic for QBCore and ESX: cinematic tuning menu, quotes, real work orders, part wear, OBD scanner, towing and police plate checks.",[19,20],7775407,"2026-11-05T16:28:51.462+00:00","2026-11-20T12:24:00.284916+00:00",{"slug":128,"title":129,"excerpt":130,"content":131,"image":132,"category":11,"author":12,"status":13,"reviewed":14,"views":133,"metaTitle":134,"metaDescription":135,"frameworks":136,"productId":22,"publishedAt":137,"updatedAt":138,"date":137},"fivem-ace-permissions-admins-groups-principals","FiveM ACE Permissions Explained: Admins, Groups and Principals","How FiveM's ACE system works, how to build a clean admin hierarchy in server.cfg, and how to check permissions safely in your own Lua resources.","\u003Cp>Most FiveM servers end up with admin access granted by copy-pasted lines nobody fully understands. That works until a staff member leaves, someone gets the wrong rank, or a resource quietly hands out command access to everyone. This guide explains how ACE permissions actually work and how to structure them so you can reason about who can do what.\u003C\u002Fp>\u003Ch3>How ACE works in one paragraph\u003C\u002Fh3>\u003Cp>ACE (Access Control Entries) has three parts. A \u003Cstrong>principal\u003C\u002Fstrong> is who is asking: a player identifier, a group, or a resource. An \u003Cstrong>object\u003C\u002Fstrong> is what they want access to, usually a command like \u003Ccode>command.kick\u003C\u002Fcode> or a custom string like \u003Ccode>myserver.vehiclespawn\u003C\u002Fcode>. An \u003Cstrong>ace\u003C\u002Fstrong> is the rule that connects them with \u003Ccode>allow\u003C\u002Fcode> or \u003Ccode>deny\u003C\u002Fcode>. Principals can inherit from other principals, which is how you build ranks.\u003C\u002Fp>\u003Cp>Four commands do almost all the work:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ccode>add_ace &lt;principal&gt; &lt;object&gt; allow|deny\u003C\u002Fcode> grants or denies an object.\u003C\u002Fli>\u003Cli>\u003Ccode>add_principal &lt;child&gt; &lt;parent&gt;\u003C\u002Fcode> makes the child inherit everything the parent has.\u003C\u002Fli>\u003Cli>\u003Ccode>remove_ace\u003C\u002Fcode> and \u003Ccode>remove_principal\u003C\u002Fcode> undo the above at runtime.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>The minimal admin setup\u003C\u002Fh3>\u003Cp>The classic setup from the default server config looks like this:\u003C\u002Fp>\u003Cpre>\u003Ccode>add_ace group.admin command allow\nadd_ace group.admin command.quit deny\nadd_principal identifier.fivem:1234567 group.admin\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>The first line allows every command for \u003Ccode>group.admin\u003C\u002Fcode>. The second explicitly denies \u003Ccode>quit\u003C\u002Fcode>, and a deny wins over an allow. The third puts a specific player into the group. Note that \u003Ccode>command\u003C\u002Fcode> on its own is a parent of every \u003Ccode>command.*\u003C\u002Fcode> object, so it is a very broad grant. That is fine for owners, not for moderators.\u003C\u002Fp>\u003Ch3>Which identifier to use\u003C\u002Fh3>\u003Cp>A principal for a player is written as \u003Ccode>identifier.&lt;type&gt;:&lt;value&gt;\u003C\u002Fcode>. Common types:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ccode>license\u003C\u002Fcode>: tied to the Rockstar license the player connects with. Always present, which makes it the most reliable choice.\u003C\u002Fli>\u003Cli>\u003Ccode>fivem\u003C\u002Fcode>: the Cfx.re forum account ID, present when the player is linked.\u003C\u002Fli>\u003Cli>\u003Ccode>discord\u003C\u002Fcode>: only present if the player has Discord running and linked.\u003C\u002Fli>\u003Cli>\u003Ccode>steam\u003C\u002Fcode>: only present if Steam is running.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Use \u003Ccode>license\u003C\u002Fcode> or \u003Ccode>fivem\u003C\u002Fcode> for staff. Granting admin through \u003Ccode>discord\u003C\u002Fcode> or \u003Ccode>steam\u003C\u002Fcode> means the permission silently disappears when the player launches without those clients, which leads to confusing support tickets. You can grab a player's identifiers from txAdmin's player panel.\u003C\u002Fp>\u003Ch3>Building a rank hierarchy\u003C\u002Fh3>\u003Cp>Instead of giving each rank its own full list, grant only the difference and use inheritance. Put this in a separate file and load it from \u003Ccode>server.cfg\u003C\u002Fcode> with \u003Ccode>exec permissions.cfg\u003C\u002Fcode> so staff changes never touch your resource list.\u003C\u002Fp>\u003Cpre>\u003Ccode># permissions.cfg\n\n# Moderator: limited commands\nadd_ace group.mod command.kick allow\nadd_ace group.mod myserver.spectate allow\nadd_ace group.mod myserver.noclip allow\n\n# Admin: everything a mod has, plus more\nadd_principal group.admin group.mod\nadd_ace group.admin command.ban allow\nadd_ace group.admin myserver.vehiclespawn allow\n\n# Owner: everything\nadd_principal group.owner group.admin\nadd_ace group.owner command allow\n\n# Staff\nadd_principal identifier.license:aaaaaaaa group.owner\nadd_principal identifier.license:bbbbbbbb group.admin\nadd_principal identifier.license:cccccccc group.mod\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Removing someone is now one line deleted, and promoting someone is changing one word. Keep a comment with the person's name next to each identifier; in six months nobody will remember who \u003Ccode>bbbbbbbb\u003C\u002Fcode> was.\u003C\u002Fp>\u003Cp>Changes in the file apply on the next restart. You can also run \u003Ccode>add_principal\u003C\u002Fcode> or \u003Ccode>remove_principal\u003C\u002Fcode> in the live console for an immediate effect, but anything typed there is lost on restart unless you also update the file.\u003C\u002Fp>\u003Ch3>Checking permissions in your own resources\u003C\u002Fh3>\u003Cp>There are two clean ways to gate functionality in Lua.\u003C\u002Fp>\u003Ch4>Restricted commands\u003C\u002Fh4>\u003Cp>The third argument of \u003Ccode>RegisterCommand\u003C\u002Fcode> marks a command as restricted. Restricted commands require the matching \u003Ccode>command.&lt;name&gt;\u003C\u002Fcode> ace, and the server enforces it before your handler runs.\u003C\u002Fp>\u003Cpre>\u003Ccode>RegisterCommand('fixcar', function(source, args)\n    -- only reached if the player has command.fixcar\n    TriggerClientEvent('myserver:fixVehicle', source)\nend, true)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Then grant it: \u003Ccode>add_ace group.mod command.fixcar allow\u003C\u002Fcode>.\u003C\u002Fp>\u003Ch4>Custom objects with IsPlayerAceAllowed\u003C\u002Fh4>\u003Cp>For anything that is not a command, such as a menu option, a server event or a button in a NUI panel, check a custom object on the server:\u003C\u002Fp>\u003Cpre>\u003Ccode>RegisterNetEvent('myserver:spawnVehicle', function(model)\n    local src = source\n    if not IsPlayerAceAllowed(src, 'myserver.vehiclespawn') then\n        print(('Blocked vehicle spawn attempt from %s'):format(src))\n        return\n    end\n    -- spawn logic here\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Always do this check server-side. A client-side check only hides a button; anyone running a modified client can still trigger the event directly. Pick a consistent prefix for your custom objects (your server name, for example) so they are easy to grep and do not collide with other resources.\u003C\u002Fp>\u003Ch3>Resource principals\u003C\u002Fh3>\u003Cp>Resources are principals too, written as \u003Ccode>resource.&lt;name&gt;\u003C\u002Fcode>. Some resources need to run console commands themselves, for example to add principals at runtime when a framework sets a player's group. That requires a line like:\u003C\u002Fp>\u003Cpre>\u003Ccode>add_ace resource.my_admin_resource command.add_principal allow\nadd_ace resource.my_admin_resource command.remove_principal allow\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Be stingy here. Granting a resource \u003Ccode>command allow\u003C\u002Fcode> means any code in that resource, including a backdoor in a leaked or obfuscated script, can execute any console command on your server. Grant only the specific commands the resource documents as necessary.\u003C\u002Fp>\u003Ch3>Where frameworks fit in\u003C\u002Fh3>\u003Cp>Frameworks layer their own role systems on top, and they do it differently:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>QBCore\u003C\u002Fstrong> and \u003Cstrong>Qbox\u003C\u002Fstrong> rely on ACE for staff permissions. Their default configs ship with specific group names and the ace lines that go with them. Read the permission section of your framework's default \u003Ccode>server.cfg\u003C\u002Fcode> and keep those group names rather than inventing parallel ones, otherwise framework admin commands will not recognise your staff.\u003C\u002Fli>\u003Cli>\u003Cstrong>ESX\u003C\u002Fstrong> traditionally stores the player's group in the database (the \u003Ccode>users\u003C\u002Fcode> table) and checks that for its own admin features. Standalone resources on the same server may still check ACE, so staff often need both the ESX group and the ACE principal.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>When a third-party script says it needs \"admin\", check its documentation for which system it reads. Mismatches between framework groups and ACE groups are the most common reason an admin \"has permission\" in one script but not another.\u003C\u002Fp>\u003Ch3>txAdmin admins are a separate layer\u003C\u002Fh3>\u003Cp>txAdmin manages its own admin accounts and permissions through its web panel, including access to the in-game txAdmin menu. Those accounts are not ACE principals. Giving someone a txAdmin account does not let them run your restricted commands, and adding them to \u003Ccode>group.admin\u003C\u002Fcode> does not give them txAdmin panel access. Decide per staff member which layers they need and keep a short written list of both.\u003C\u002Fp>\u003Ch3>Debugging permissions\u003C\u002Fh3>\u003Cp>When someone says a command does not work for them, check in this order:\u003C\u002Fp>\u003Col>\u003Cli>Confirm the identifier. Copy it from txAdmin rather than from memory, and check the prefix (\u003Ccode>license:\u003C\u002Fcode> vs \u003Ccode>license2:\u003C\u002Fcode> are different).\u003C\u002Fli>\u003Cli>Run \u003Ccode>list_principals\u003C\u002Fcode> in the server console to see the inheritance tree that is actually loaded.\u003C\u002Fli>\u003Cli>Run \u003Ccode>list_aces\u003C\u002Fcode> to see every ace that is active, including ones added by resources.\u003C\u002Fli>\u003Cli>Use \u003Ccode>test_ace\u003C\u002Fcode> with the principal and object to see whether the server resolves it as allowed.\u003C\u002Fli>\u003Cli>Check that your \u003Ccode>exec permissions.cfg\u003C\u002Fcode> line actually runs; a typo in the file name fails quietly in a busy console log.\u003C\u002Fli>\u003C\u002Fol>\u003Ch3>Common mistakes\u003C\u002Fh3>\u003Cul>\u003Cli>Granting anything to \u003Ccode>builtin.everyone\u003C\u002Fcode> or a default group that every player inherits. Audit for this after installing any new resource that edits your config.\u003C\u002Fli>\u003Cli>Giving moderators \u003Ccode>command allow\u003C\u002Fcode> because it was faster than listing commands.\u003C\u002Fli>\u003Cli>Using Discord or Steam identifiers for staff ranks.\u003C\u002Fli>\u003Cli>Checking permissions only on the client.\u003C\u002Fli>\u003Cli>Leaving former staff in the file because nobody owns the list.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Practical takeaway\u003C\u002Fh3>\u003Cp>Move all permissions into one \u003Ccode>permissions.cfg\u003C\u002Fcode>, build ranks with inheritance so each group only lists what it adds, use \u003Ccode>license\u003C\u002Fcode> or \u003Ccode>fivem\u003C\u002Fcode> identifiers, and gate every sensitive server event with \u003Ccode>IsPlayerAceAllowed\u003C\u002Fcode>. Review the file whenever staff changes and after every new resource install. For the full list of console commands and natives, the official documentation at \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\" target=\"_blank\" rel=\"noopener noreferrer\">docs.fivem.net\u003C\u002Fa> is the reference to keep open.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-ace-permissions-admins-groups-principals.webp?v=mvxzvv0x",25,"FiveM ACE Permissions: Admin Groups Done Right","Set up FiveM ACE permissions properly: principals, groups, inheritance, restricted commands, IsPlayerAceAllowed and the mistakes that leak admin access.",[19,20,21,49],"2026-11-02T11:52:06.7+00:00","2026-11-20T09:23:04.321015+00:00",{"slug":140,"title":141,"excerpt":142,"content":143,"image":144,"category":11,"author":12,"status":13,"reviewed":14,"views":145,"metaTitle":146,"metaDescription":147,"frameworks":148,"productId":22,"publishedAt":149,"updatedAt":150,"date":149},"fivem-server-backups-what-to-save-and-how-to-restore","FiveM Server Backups: What to Save, How Often, and How to Restore","A practical backup plan for FiveM servers: which files and databases matter, how to automate dumps, where to store them, and how to test a restore.","\u003Cp>Most FiveM servers do not lose data because of hackers. They lose it because of a bad update, a dropped table, a dead VPS disk or an admin who ran the wrong SQL query at 2 a.m. A working backup plan turns those events into a 15-minute inconvenience instead of a wipe. This guide covers what to back up, how to automate it, and how to prove your backups actually restore.\u003C\u002Fp>\u003Ch3>What actually needs backing up\u003C\u002Fh3>\u003Cp>A FiveM server is a few distinct pieces, and each one fails in a different way. Treat them separately.\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>The database.\u003C\u002Fstrong> This is the only thing you cannot rebuild. Characters, money, inventories, vehicles, houses, job grades and bans all live in MariaDB or MySQL, whether you run QBCore, ESX or Qbox. If you only back up one thing, back up this.\u003C\u002Fli>\u003Cli>\u003Cstrong>Your server-data folder.\u003C\u002Fstrong> The \u003Ccode>resources\u003C\u002Fcode> directory with every script you installed and every config you edited, plus \u003Ccode>server.cfg\u003C\u002Fcode>. You can technically re-download scripts, but rebuilding months of config tweaks from memory is not realistic.\u003C\u002Fli>\u003Cli>\u003Cstrong>Secrets.\u003C\u002Fstrong> Your license key, database connection string, RCON password if you use one, and any API keys or webhooks. These usually sit inside \u003Ccode>server.cfg\u003C\u002Fcode> or a file it executes.\u003C\u002Fli>\u003Cli>\u003Cstrong>txAdmin data.\u003C\u002Fstrong> The \u003Ccode>txData\u003C\u002Fcode> folder holds txAdmin's own configuration, admin accounts and its player database (warnings, bans, notes). Its location depends on how you launched txAdmin, so find it once and write the path down.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>What you can skip: the \u003Ccode>cache\u003C\u002Fcode> folder inside your server-data directory. FXServer regenerates it, it is large, and restoring a stale cache can cause more confusion than it solves. The server artifacts themselves are also replaceable, but note which artifact build you were running so you can return to it if an update breaks something.\u003C\u002Fp>\u003Ch3>Database dumps done properly\u003C\u002Fh3>\u003Cp>Copying the raw MariaDB data directory while the database is running is a common mistake and can give you a corrupted copy. Use a logical dump instead. For InnoDB tables, which is what modern framework schemas use, \u003Ccode>--single-transaction\u003C\u002Fcode> gives you a consistent snapshot without locking the server while players are online.\u003C\u002Fp>\u003Cpre>\u003Ccode>mysqldump --single-transaction --routines --triggers \\\n  --default-character-set=utf8mb4 fivem | gzip &gt; fivem_2024-01-01.sql.gz\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>A few details that matter:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Character set.\u003C\u002Fstrong> Player names and item labels often contain non-ASCII characters. Dumping with \u003Ccode>utf8mb4\u003C\u002Fcode> avoids mangled names after a restore.\u003C\u002Fli>\u003Cli>\u003Cstrong>Credentials.\u003C\u002Fstrong> Do not put the database password on the command line where it ends up in shell history. On Linux, put it in a \u003Ccode>~\u002F.my.cnf\u003C\u002Fcode> file owned by the backup user with permissions set to \u003Ccode>600\u003C\u002Fcode>.\u003C\u002Fli>\u003Cli>\u003Cstrong>Newer MariaDB.\u003C\u002Fstrong> Recent MariaDB versions ship the same tool as \u003Ccode>mariadb-dump\u003C\u002Fcode>. The flags above work the same way.\u003C\u002Fli>\u003Cli>\u003Cstrong>Windows hosts.\u003C\u002Fstrong> Exporting from HeidiSQL is fine for a one-off, but it is manual and gets forgotten. Script it with the dump tool and Task Scheduler instead.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>A simple automated backup script\u003C\u002Fh3>\u003Cp>Here is a minimal Linux script that dumps the database, archives the server-data folder without the cache, and deletes local copies older than two weeks. Adjust the paths and database name to your setup.\u003C\u002Fp>\u003Cpre>\u003Ccode>#!\u002Fbin\u002Fbash\nset -euo pipefail\n\nSTAMP=$(date +%Y-%m-%d_%H%M)\nDEST=\u002Fbackups\u002Ffivem\nDB=fivem\nDATA=\u002Fhome\u002Ffivem\u002Fserver-data\n\nmkdir -p \"$DEST\"\n\nmysqldump --single-transaction --routines --triggers \\\n  --default-character-set=utf8mb4 \"$DB\" | gzip &gt; \"$DEST\u002Fdb_$STAMP.sql.gz\"\n\ntar --exclude='.\u002Fcache' -czf \"$DEST\u002Ffiles_$STAMP.tar.gz\" -C \"$DATA\" .\n\nfind \"$DEST\" -type f -mtime +14 -delete\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Schedule it with cron. Running it shortly before your daily txAdmin scheduled restart works well, because the restart then starts from the state you just saved:\u003C\u002Fp>\u003Cpre>\u003Ccode>30 4 * * * \u002Fhome\u002Ffivem\u002Fbackup.sh\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Add the \u003Ccode>txData\u003C\u002Fcode> folder to the \u003Ccode>tar\u003C\u002Fcode> step, or to a second archive, once you know its path.\u003C\u002Fp>\u003Ch3>Keep secrets out of shared copies\u003C\u002Fh3>\u003Cp>Backups get copied to more places than your live server: cloud storage, a co-owner's laptop, a Git repository. Each copy is a chance to leak your license key or database password.\u003C\u002Fp>\u003Cp>A clean pattern is to move secrets into their own file and load it from \u003Ccode>server.cfg\u003C\u002Fcode>:\u003C\u002Fp>\u003Cpre>\u003Ccode># server.cfg\nexec secrets.cfg\nensure oxmysql\nensure qb-core\n\n# secrets.cfg (never commit this file)\nsv_licenseKey \"your_key_here\"\nset mysql_connection_string \"mysql:\u002F\u002Fuser:pass@localhost\u002Ffivem\"\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Now \u003Ccode>server.cfg\u003C\u002Fcode> and your resources can go into a private Git repository with \u003Ccode>secrets.cfg\u003C\u002Fcode> listed in \u003Ccode>.gitignore\u003C\u002Fcode>, and you back up the secrets file separately, encrypted. Git also gives you something archives do not: a history of every config change, so you can see exactly what changed before things broke.\u003C\u002Fp>\u003Ch3>Store copies somewhere else\u003C\u002Fh3>\u003Cp>A backup on the same disk as the server protects you from a bad query, not from a dead machine or a host that suspends your account. Follow the usual 3-2-1 idea, scaled to a game server:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Local copy\u003C\u002Fstrong> on the server for fast restores after a mistake.\u003C\u002Fli>\u003Cli>\u003Cstrong>Offsite copy\u003C\u002Fstrong> pushed automatically to object storage or a different provider. Tools like \u003Ccode>rclone\u003C\u002Fcode> or \u003Ccode>restic\u003C\u002Fcode> handle this and can encrypt before upload.\u003C\u002Fli>\u003Cli>\u003Cstrong>Occasional cold copy\u003C\u002Fstrong>, for example a weekly download to a machine you control.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Restrict who can read the offsite bucket. A database dump contains every player identifier you store, so treat it with the same care as the live database.\u003C\u002Fp>\u003Ch3>When to take extra backups\u003C\u002Fh3>\u003Cp>Daily automated backups cover normal operation. Take a manual one, and label it clearly, before any of these:\u003C\u002Fp>\u003Col>\u003Cli>Updating server artifacts.\u003C\u002Fli>\u003Cli>Updating your framework core or a large resource such as an inventory, housing or garage system.\u003C\u002Fli>\u003Cli>Running any SQL file that ships with a new script, especially ones that \u003Ccode>ALTER\u003C\u002Fcode> or \u003Ccode>DROP\u003C\u002Fcode> tables.\u003C\u002Fli>\u003Cli>Bulk admin actions: economy resets, mass item removals, character wipes.\u003C\u002Fli>\u003Cli>Migrating hosts or moving between frameworks.\u003C\u002Fli>\u003C\u002Fol>\u003Cp>A dump named \u003Ccode>db_before-inventory-update.sql.gz\u003C\u002Fcode> is worth far more in a crisis than the fifth file with a timestamp you have to guess at.\u003C\u002Fp>\u003Ch3>Test the restore, not just the backup\u003C\u002Fh3>\u003Cp>An untested backup is a guess. Do a restore drill at least once, and again whenever you change the script. The safest way is to restore into a separate database so you never touch live data:\u003C\u002Fp>\u003Cpre>\u003Ccode>mysql -e \"CREATE DATABASE fivem_restore CHARACTER SET utf8mb4\"\ngunzip -c \u002Fbackups\u002Ffivem\u002Fdb_2024-01-01_0430.sql.gz | mysql fivem_restore\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Then check it like you mean it:\u003C\u002Fp>\u003Cul>\u003Cli>Row counts on your players or users table roughly match production.\u003C\u002Fli>\u003Cli>A known character's money, inventory and owned vehicles look correct.\u003C\u002Fli>\u003Cli>Names with special characters display properly.\u003C\u002Fli>\u003Cli>Extract the file archive on a test machine, point a copy of the server at \u003Ccode>fivem_restore\u003C\u002Fcode> using a separate license key, and confirm it boots and you can log in.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Time the whole process. Knowing that a full restore takes 20 minutes lets you make calm decisions during an outage instead of improvising. Write the steps in a short document your other admins can follow without you.\u003C\u002Fp>\u003Ch3>Rolling back a single mistake\u003C\u002Fh3>\u003Cp>Not every problem needs a full restore. If one player's inventory was wiped by a bug, restoring the whole database would roll back everyone else's progress since the last dump. Instead, load the backup into \u003Ccode>fivem_restore\u003C\u002Fcode>, query the affected rows there, and copy only those values back to production. This is another reason to keep restoring to a separate database as your default habit.\u003C\u002Fp>\u003Ch3>Practical takeaway\u003C\u002Fh3>\u003Cul>\u003Cli>Back up the database daily with a consistent dump, compressed and dated.\u003C\u002Fli>\u003Cli>Archive server-data and \u003Ccode>txData\u003C\u002Fcode>, excluding \u003Ccode>cache\u003C\u002Fcode>.\u003C\u002Fli>\u003Cli>Move secrets into their own file and keep them out of Git and shared copies.\u003C\u002Fli>\u003Cli>Push encrypted copies offsite automatically.\u003C\u002Fli>\u003Cli>Take a labelled manual backup before every update or SQL change.\u003C\u002Fli>\u003Cli>Run a restore drill into a separate database and write the steps down.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>If you are still setting up the server itself, the official \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\u002Fdocs\u002Fserver-manual\u002Fsetting-up-a-server\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">FiveM server manual\u003C\u002Fa> covers artifacts and txAdmin. Put the backup script in place on day one, before your players have anything to lose.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-server-backups-what-to-save-and-how-to-restore.webp?v=mvwiyj5o",26,"FiveM Server Backups: A Practical Restore-First Guide","Learn what to back up on a FiveM server, how to automate database dumps and file archives, store them offsite, and actually test a restore.",[19,20,21],"2026-10-31T14:56:54.725+00:00","2026-11-19T20:49:25.981484+00:00",{"slug":152,"title":153,"excerpt":154,"content":155,"image":156,"category":11,"author":12,"status":13,"reviewed":14,"views":157,"metaTitle":158,"metaDescription":159,"frameworks":160,"productId":22,"publishedAt":161,"updatedAt":162,"date":161},"fivem-state-bags-lua-syncing-data","FiveM State Bags in Lua: Syncing Data Without Event Spam","How to use entity, player and global state bags in FiveM Lua to replicate data cleanly, react to changes, and keep the server in charge.","\u003Cp>Many FiveM scripts still sync shared data the old way: the server fires \u003Ccode>TriggerClientEvent\u003C\u002Fcode> to everyone whenever something changes, and every client keeps its own copy in a table that slowly drifts out of date. State bags solve most of that. This article covers how they work in Lua, when to use each type, how to react to changes, and the mistakes that cause desync and exploits.\u003C\u002Fp>\u003Ch3>What a state bag actually is\u003C\u002Fh3>\u003Cp>A state bag is a key-value store attached to something the server knows about: a networked entity, a player, or the server as a whole. When a value is set with replication enabled, the new value is sent to clients automatically, including clients that join or stream the entity in later. You stop managing \"who needs to know\" yourself.\u003C\u002Fp>\u003Cp>There are three entry points you will use in practice:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Entity state\u003C\u002Fstrong>: \u003Ccode>Entity(entity).state\u003C\u002Fcode>, attached to a networked entity such as a vehicle, ped or object.\u003C\u002Fli>\u003Cli>\u003Cstrong>Player state\u003C\u002Fstrong>: \u003Ccode>Player(source).state\u003C\u002Fcode> on the server, \u003Ccode>LocalPlayer.state\u003C\u002Fcode> on the client for the local player.\u003C\u002Fli>\u003Cli>\u003Cstrong>Global state\u003C\u002Fstrong>: \u003Ccode>GlobalState\u003C\u002Fcode>, one bag shared by the whole server.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>The official reference lives in the \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\u002Fdocs\u002Fscripting-manual\u002Fnetworking\u002Fstate-bags\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">FiveM state bags documentation\u003C\u002Fa>. Read it once; the rest of this article focuses on how to use them well.\u003C\u002Fp>\u003Ch3>Choosing the right bag\u003C\u002Fh3>\u003Cp>Picking the wrong bag type is the most common design mistake. A quick rule set:\u003C\u002Fp>\u003Cul>\u003Cli>Data that belongs to a physical thing in the world (a vehicle is locked, a crate has been looted, a door object is broken) goes in \u003Cstrong>entity state\u003C\u002Fstrong>. It lives and dies with the entity.\u003C\u002Fli>\u003Cli>Data about a person for the current session (on duty, current job grade shown on a HUD, a temporary status) goes in \u003Cstrong>player state\u003C\u002Fstrong>. It is cleared when the player drops.\u003C\u002Fli>\u003Cli>Server-wide values everyone needs (a heist cooldown flag, the current event phase, whether a store is open) go in \u003Cstrong>global state\u003C\u002Fstrong>.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>None of these are persistent. A restart wipes them. If something must survive a restart, the database is the source of truth and the state bag is only a live mirror of it.\u003C\u002Fp>\u003Ch3>Setting values from the server\u003C\u002Fh3>\u003Cp>The safest pattern is: the client asks, the server validates, the server writes the state. Here is a vehicle lock toggle that follows it.\u003C\u002Fp>\u003Cpre>\u003Ccode>-- server.lua\nRegisterNetEvent('myres:toggleLock', function(netId)\n    local src = source\n    local veh = NetworkGetEntityFromNetworkId(netId)\n    if veh == 0 or not DoesEntityExist(veh) then return end\n\n    local ped = GetPlayerPed(src)\n    if #(GetEntityCoords(ped) - GetEntityCoords(veh)) &gt; 10.0 then return end\n\n    -- add your own ownership \u002F key check here\n\n    local state = Entity(veh).state\n    state:set('locked', not state.locked, true)\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>The third argument to \u003Ccode>set\u003C\u002Fcode> is the replicate flag. With \u003Ccode>true\u003C\u002Fcode>, clients receive the value. With \u003Ccode>false\u003C\u002Fcode>, it stays on the server, which is useful for internal bookkeeping you do not want players to read, such as who spawned the entity or an anti-abuse counter.\u003C\u002Fp>\u003Cp>Player and global state work the same way:\u003C\u002Fp>\u003Cpre>\u003Ccode>-- server.lua\nPlayer(src).state:set('onDuty', true, true)\nGlobalState.bankCooldown = true\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Assigning directly to a key, as in the \u003Ccode>GlobalState\u003C\u002Fcode> line, replicates by default. Using \u003Ccode>set\u003C\u002Fcode> makes the intent explicit, which is easier to review later.\u003C\u002Fp>\u003Ch3>Reading and reacting on the client\u003C\u002Fh3>\u003Cp>Reading is a plain table lookup: \u003Ccode>Entity(veh).state.locked\u003C\u002Fcode> or \u003Ccode>LocalPlayer.state.onDuty\u003C\u002Fcode>. That is fine for one-off checks, for example when the player presses a key.\u003C\u002Fp>\u003Cp>For anything that should apply the moment a value changes, register a change handler instead of polling in a loop:\u003C\u002Fp>\u003Cpre>\u003Ccode>-- client.lua\nAddStateBagChangeHandler('locked', nil, function(bagName, key, value)\n    local veh = GetEntityFromStateBagName(bagName)\n    if veh == 0 then return end\n    SetVehicleDoorsLocked(veh, value and 2 or 1)\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>The first argument filters by key, the second by bag name. Passing \u003Ccode>nil\u003C\u002Fcode> for the bag means \"any bag\", so this one handler covers every vehicle on the server. That is far cheaper than a thread that loops through nearby vehicles every frame checking a table.\u003C\u002Fp>\u003Ch4>Two details that bite people\u003C\u002Fh4>\u003Cul>\u003Cli>\u003Cstrong>Use the \u003Ccode>value\u003C\u002Fcode> argument.\u003C\u002Fstrong> Inside the handler, the bag may still hold the old value. Reading \u003Ccode>Entity(veh).state.locked\u003C\u002Fcode> there can give you the previous state. Always trust the parameter.\u003C\u002Fli>\u003Cli>\u003Cstrong>The entity may not exist locally.\u003C\u002Fstrong> If the vehicle is out of scope for this client, \u003Ccode>GetEntityFromStateBagName\u003C\u002Fcode> returns \u003Ccode>0\u003C\u002Fcode>. When that vehicle streams in later, the handler will not necessarily fire again for a value that did not change. Apply the current state wherever you first interact with the entity too, for example when a player enters the vehicle or targets it.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Never let the client be the authority\u003C\u002Fh3>\u003Cp>Clients can write to some state bags themselves, such as their own player bag and entities they control. That is convenient for purely cosmetic data, but it means any value a client can write, a cheater can write too.\u003C\u002Fp>\u003Cp>Treat it this way:\u003C\u002Fp>\u003Cul>\u003Cli>Anything that grants money, items, access or job permissions must be set by the server only, and the server must decide based on its own data, never on a value it read back from a client-writable bag.\u003C\u002Fli>\u003Cli>If you allow client writes for something harmless (a selected walk style, a UI preference), keep it isolated under clearly named keys and never use those keys in server-side logic that matters.\u003C\u002Fli>\u003Cli>On the server, a change handler receives a \u003Ccode>replicated\u003C\u002Fcode> flag indicating the change came from a client. You can use it to log or revert writes to keys that only the server should touch.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>A quick server-side guard looks like this:\u003C\u002Fp>\u003Cpre>\u003Ccode>-- server.lua\nlocal protectedKeys = { onDuty = true, jobGrade = true }\n\nAddStateBagChangeHandler(nil, nil, function(bagName, key, value, _, replicated)\n    if replicated and protectedKeys[key] then\n        print(('[myres] client tried to set %s on %s'):format(key, bagName))\n    end\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Logging is a starting point. The real protection is that your server code never reads these keys as proof of anything; it reads its own records.\u003C\u002Fp>\u003Ch3>Keep payloads small\u003C\u002Fh3>\u003Cp>Every replicated change sends the whole value for that key, not a diff. If you store a 200-entry table in one key and change one field, the full table goes over the network to every client that should receive it.\u003C\u002Fp>\u003Cul>\u003Cli>Split large structures into several keys so a change only resends what moved.\u003C\u002Fli>\u003Cli>Store ids and short flags, not full item lists or descriptions. Clients can look up static data from a shared config file.\u003C\u002Fli>\u003Cli>Do not update a replicated key every frame or every few hundred milliseconds. If a value changes that often, it is probably something each client should compute locally.\u003C\u002Fli>\u003Cli>Use non-replicated keys for server-only data instead of keeping a separate Lua table that can fall out of sync with the entity.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Cleanup and lifecycle\u003C\u002Fh3>\u003Cp>Entity state disappears when the entity is deleted. Player state disappears when the player drops. That is usually what you want, but it has consequences:\u003C\u002Fp>\u003Cul>\u003Cli>If your script tracks something important on an entity, also save it to the database when it matters, or it is gone with the entity.\u003C\u002Fli>\u003Cli>Global state persists until the server restarts or you clear it. Reset keys to \u003Ccode>nil\u003C\u002Fcode> when an event ends so stale flags do not confuse other resources.\u003C\u002Fli>\u003Cli>Restarting your resource does not clear state that was set on entities or globally. On resource start, either re-read existing state or reset your keys deliberately.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Debugging state bags\u003C\u002Fh3>\u003Cp>When something does not sync, check in this order:\u003C\u002Fp>\u003Col>\u003Cli>Was the value set with replication enabled? A \u003Ccode>false\u003C\u002Fcode> flag is the usual culprit.\u003C\u002Fli>\u003Cli>Is the entity actually networked? Local-only entities have no network id, and their state will not reach other clients.\u003C\u002Fli>\u003Cli>Is the client in scope? Test with two players standing next to each other before testing across the map.\u003C\u002Fli>\u003Cli>Is the handler key spelled the same as the key you set? A key filter is an exact match.\u003C\u002Fli>\u003Cli>Add a temporary handler with \u003Ccode>nil, nil\u003C\u002Fcode> filters that prints every change, then remove it once you find the problem. It is noisy but shows exactly what is replicating.\u003C\u002Fli>\u003C\u002Fol>\u003Ch3>Practical takeaway\u003C\u002Fh3>\u003Cul>\u003Cli>Use entity state for things in the world, player state for session data, global state for server-wide flags, and the database for anything that must persist.\u003C\u002Fli>\u003Cli>Let the server write anything that matters; clients request, the server validates and sets.\u003C\u002Fli>\u003Cli>React with \u003Ccode>AddStateBagChangeHandler\u003C\u002Fcode> instead of polling loops, and use the handler's \u003Ccode>value\u003C\u002Fcode> argument.\u003C\u002Fli>\u003Cli>Handle entities that are not yet streamed in by also applying state when you first interact with them.\u003C\u002Fli>\u003Cli>Keep each key small and change it rarely. If you would not send it in an event every second, do not put it in a replicated key that changes every second.\u003C\u002Fli>\u003C\u002Ful>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Ffivem-state-bags-lua-syncing-data.webp?v=mvu3z3hf",33,"FiveM State Bags in Lua: A Practical Guide","Learn how to use FiveM state bags in Lua: entity, player and global state, change handlers, server authority, payload size and common pitfalls.",[19,20,21,49],"2026-10-29T22:22:00.491+00:00","2026-11-21T08:13:02.187624+00:00",{"slug":164,"title":165,"excerpt":166,"content":167,"image":168,"category":11,"author":12,"status":13,"reviewed":14,"views":169,"metaTitle":170,"metaDescription":171,"frameworks":172,"productId":22,"publishedAt":173,"updatedAt":174,"date":173},"securing-fivem-server-events-never-trust-the-client","Never Trust the Client: Securing FiveM Server Events in Lua","Most money exploits on FiveM servers come from server events that trust whatever the client sends. Here is how to validate who, where, when and how much.","\u003Cp>Most money and item exploits on FiveM servers do not need a clever cheat. They need one server event that accepts an amount, an item name or a coordinate from the client and acts on it. This article shows how to write server events that stay safe even when the person calling them is running an executor.\u003C\u002Fp>\u003Ch3>The threat model: every client message is a suggestion\u003C\u002Fh3>\u003Cp>Anything that runs on the client can be read, modified or replayed. That includes your client Lua, your NUI JavaScript and every \u003Ccode>TriggerServerEvent\u003C\u002Fcode> call. A cheater does not need your source code to find your events. They can watch what fires while they play your job normally, then call the same event with different arguments, a thousand times, from anywhere on the map.\u003C\u002Fp>\u003Cp>So the rule is simple: the client may \u003Cem>ask\u003C\u002Fem> for something to happen. The server decides whether it happens and how much it is worth. Event names do not protect you either. Renaming \u003Ccode>job:pay\u003C\u002Fcode> to something random only slows down someone who is already watching traffic.\u003C\u002Fp>\u003Ch3>The classic exploit: a payout event with an amount\u003C\u002Fh3>\u003Cp>This pattern still shows up in free and paid resources:\u003C\u002Fp>\u003Cpre>\u003Ccode>-- client\nTriggerServerEvent('garbage:pay', 5000)\n\n-- server\nRegisterNetEvent('garbage:pay', function(amount)\n    local Player = QBCore.Functions.GetPlayer(source)\n    Player.Functions.AddMoney('cash', amount)\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Three problems in four lines. The client chooses the amount. Nothing checks whether the player is doing the job, is near a bin, or has already been paid for this bin. And nothing limits how often the event can be called. Fix any one of these and the exploit gets harder; fix all of them and it stops being worth the effort.\u003C\u002Fp>\u003Ch3>Move the state to the server\u003C\u002Fh3>\u003Cp>The most effective change is not a check, it is a design change. The server should know what the player is doing, so the client never has to tell it. When a player starts a job, the server creates the run, picks the route from config and sends the client only what it needs to display.\u003C\u002Fp>\u003Cpre>\u003Ccode>local activeRuns = {}\n\nRegisterNetEvent('garbage:server:startRun', function()\n    local src = source\n    if activeRuns[src] then return end\n    if not hasJob(src, 'garbage') then return end\n\n    local route = {}\n    for i, coords in ipairs(Config.Bins) do\n        route[i] = { coords = coords, done = false }\n    end\n\n    activeRuns[src] = { route = route, startedAt = os.time() }\n    TriggerClientEvent('garbage:client:setRoute', src, Config.Bins)\nend)\n\nAddEventHandler('playerDropped', function()\n    activeRuns[source] = nil\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Now the client reports progress with an index into a list the server owns, not with money or coordinates. The server can look up the real bin position and the real reward itself. Always clean up on \u003Ccode>playerDropped\u003C\u002Fcode>, or your tables leak memory and a reconnecting player may inherit stale state.\u003C\u002Fp>\u003Ch3>Validate who, where, when and what\u003C\u002Fh3>\u003Cp>With state on the server, the completion handler can run four cheap checks before paying anything.\u003C\u002Fp>\u003Cpre>\u003Ccode>local PAY_PER_BAG = 150\nlocal MAX_DISTANCE = 10.0\nlocal COOLDOWN_MS = 8000\nlocal lastDrop = {}\n\nRegisterNetEvent('garbage:server:dropBag', function(binIndex)\n    local src = source\n\n    -- who: is this player actually on a run?\n    local run = activeRuns[src]\n    if not run then return end\n\n    -- what: is the argument the type and range we expect?\n    if type(binIndex) ~= 'number' or binIndex % 1 ~= 0 then return end\n    local bin = run.route[binIndex]\n    if not bin or bin.done then return end\n\n    -- when: rate limit per player\n    local now = GetGameTimer()\n    if lastDrop[src] and now - lastDrop[src] &lt; COOLDOWN_MS then return end\n\n    -- where: is the ped near the bin the server knows about?\n    local ped = GetPlayerPed(src)\n    local dist = #(GetEntityCoords(ped) - bin.coords)\n    if dist &gt; MAX_DISTANCE then\n        print(('[garbage] %s (%d) dropped bag %.1fm from bin %d'):format(\n            GetPlayerName(src), src, dist, binIndex))\n        return\n    end\n\n    bin.done = true\n    lastDrop[src] = now\n    addMoney(src, PAY_PER_BAG)\nend)\u003C\u002Fcode>\u003C\u002Fpre>\u003Ch4>Who\u003C\u002Fh4>\u003Cp>Check the job and the active task, not just the job. A player with the garbage job who never started a run should get nothing. For police or EMS actions, check the job and duty status every time, because players can be fired or go off duty between two events.\u003C\u002Fp>\u003Ch4>What\u003C\u002Fh4>\u003Cp>Check types and ranges for every argument. Lua will happily index a table with a string, a negative number or a huge float. If you expect an integer between 1 and the route length, test exactly that. For strings such as item names, look them up in a server-side whitelist instead of passing them straight to your inventory.\u003C\u002Fp>\u003Ch4>When\u003C\u002Fh4>\u003Cp>A per-player cooldown kills replay spam. Choose a value just below the fastest legitimate time to complete the action. Server-side \u003Ccode>GetGameTimer\u003C\u002Fcode> is fine for this; use \u003Ccode>os.time\u003C\u002Fcode> when you need wall-clock values for logs or database rows.\u003C\u002Fp>\u003Ch4>Where\u003C\u002Fh4>\u003Cp>With OneSync enabled, the server can read a player's ped position with \u003Ccode>GetPlayerPed\u003C\u002Fcode> and \u003Ccode>GetEntityCoords\u003C\u002Fcode>. Use the coordinates you stored, never coordinates the client sent. Give a generous tolerance; network latency and animations mean a strict 2 metre check will punish honest players.\u003C\u002Fp>\u003Ch3>Capture source before anything async\u003C\u002Fh3>\u003Cp>\u003Ccode>source\u003C\u002Fcode> is a global that changes as other events fire. If your handler waits on a database query, a \u003Ccode>Wait\u003C\u002Fcode> or a callback, \u003Ccode>source\u003C\u002Fcode> may point to a different player when it resumes. Always assign \u003Ccode>local src = source\u003C\u002Fcode> on the first line and use \u003Ccode>src\u003C\u002Fcode> everywhere after that. This bug is subtle because it only shows up under load, which is exactly when you are least able to debug it.\u003C\u002Fp>\u003Ch3>Keep the money call in one place\u003C\u002Fh3>\u003Cp>Wrap framework calls in a single helper so every payout goes through the same path, and so you can add logging later without touching every job.\u003C\u002Fp>\u003Cpre>\u003Ccode>local function addMoney(src, amount)\n    -- QBCore\n    local Player = QBCore.Functions.GetPlayer(src)\n    if Player then\n        Player.Functions.AddMoney('cash', amount, 'garbage-job')\n    end\n\n    -- ESX equivalent:\n    -- local xPlayer = ESX.GetPlayerFromId(src)\n    -- if xPlayer then xPlayer.addMoney(amount) end\nend\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>On Qbox, use the equivalent function from your core. The point is that \u003Ccode>amount\u003C\u002Fcode> comes from your config, never from an event argument.\u003C\u002Fp>\u003Ch3>State bags are input too\u003C\u002Fh3>\u003Cp>State bags are a clean way to share data like duty status or a carried object. But a client can set replicated values on its own player state, for example through \u003Ccode>LocalPlayer.state\u003C\u002Fcode>. That means a value like \u003Ccode>onDuty\u003C\u002Fcode> or \u003Ccode>isCuffed\u003C\u002Fcode> that the client can write is just another message from the client.\u003C\u002Fp>\u003Cp>Treat it accordingly. Set authoritative values from the server with \u003Ccode>Player(src).state:set(key, value, true)\u003C\u002Fcode>, and when the server reads a state bag to make a decision, make sure that key is only ever written by server code. If you need a client to report something, use an event you validate, then let the server write the state.\u003C\u002Fp>\u003Ch3>NUI callbacks follow the same rule\u003C\u002Fh3>\u003Cp>Your NUI posts to a client-side callback, which then usually triggers a server event. A player can open devtools on the NUI page or call the client callback directly, so prices, quantities and item names shown in your UI mean nothing on the server. Re-read them from config, recompute totals and check stock on the server before charging or giving anything.\u003C\u002Fp>\u003Ch3>Log first, punish carefully\u003C\u002Fh3>\u003Cp>It is tempting to call \u003Ccode>DropPlayer\u003C\u002Fcode> the moment a check fails. Do not do that for distance or timing checks. Desync, lag spikes and teleporting admins will trigger them on honest players. A better approach:\u003C\u002Fp>\u003Cul>\u003Cli>Silently reject the request and return.\u003C\u002Fli>\u003Cli>Log the player, the check that failed and the values involved.\u003C\u002Fli>\u003Cli>Count failures per player and only alert staff after a threshold.\u003C\u002Fli>\u003Cli>Reserve automatic kicks for impossible inputs, such as a negative quantity or a bin index that does not exist.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Logs also tell you when your tolerances are wrong. If half your legitimate players fail the distance check at one bin, the config coordinates are off, not the players.\u003C\u002Fp>\u003Ch3>Checklist for every server event\u003C\u002Fh3>\u003Col>\u003Cli>Is \u003Ccode>source\u003C\u002Fcode> captured into a local on the first line?\u003C\u002Fli>\u003Cli>Does the server already know what this player is doing, or is it taking the client's word?\u003C\u002Fli>\u003Cli>Are all arguments type-checked and range-checked?\u003C\u002Fli>\u003Cli>Are amounts, prices and rewards read from server config?\u003C\u002Fli>\u003Cli>Is the player's job, duty or ownership checked on every call?\u003C\u002Fli>\u003Cli>Is the position checked against server-known coordinates?\u003C\u002Fli>\u003Cli>Is there a cooldown or a one-time flag preventing replay?\u003C\u002Fli>\u003Cli>Is per-player state cleared on \u003Ccode>playerDropped\u003C\u002Fcode>?\u003C\u002Fli>\u003Cli>Are failures logged with enough detail to tune thresholds?\u003C\u002Fli>\u003C\u002Fol>\u003Ch3>Takeaway\u003C\u002Fh3>\u003Cp>Secure events are mostly about where the data lives. Keep tasks, routes, rewards and permissions on the server, let the client send only intents and indexes, and validate who, what, when and where before touching money or items. Read the \u003Ca href=\"https:\u002F\u002Fdocs.fivem.net\u002Fdocs\u002Fscripting-manual\u002Fworking-with-events\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">FiveM events documentation\u003C\u002Fa> for the details on net events, then run the checklist above against every \u003Ccode>RegisterNetEvent\u003C\u002Fcode> in your resources. If a handler fails two or more items, fix it before you ship it.\u003C\u002Fp>","https:\u002F\u002Fkhkzoorftujlbvlkwxbi.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-covers\u002Fsecuring-fivem-server-events-never-trust-the-client.webp?v=mvu3qn0y",29,"Securing FiveM Server Events: Never Trust the Client","A practical guide for FiveM script authors: validate net events on the server, keep job state server-side, rate limit, check distance and stop payout exploits.",[19,20,21],"2026-10-29T22:11:13.362+00:00","2026-11-21T03:11:16.418058+00:00",{"slug":176,"title":177,"excerpt":178,"content":179,"image":180,"category":11,"author":12,"status":13,"reviewed":14,"views":181,"metaTitle":182,"metaDescription":183,"frameworks":184,"productId":185,"publishedAt":186,"updatedAt":187,"date":186},"fivem-resource-optimization","FiveM Resource Optimization: Writing Scripts That Cost Nothing","The difference between a 0.00ms resource and a 2ms one is rarely clever code. It is not running code at all when nothing is happening. Here are the patterns that get you there.","\u003Cp>A FiveM resource that costs 0.00ms and one that costs 2ms usually do the same work. The difference is when they do it. This guide covers the patterns that separate the two, whether you are writing your own scripts or judging someone else's before you buy.\u003C\u002Fp>\u003Ch3>The per-frame loop is the whole problem\u003C\u002Fh3>\u003Cp>Most expensive resources contain some version of this:\u003C\u002Fp>\u003Cp>\u003Ccode>while true do Wait(0) local pos = GetEntityCoords(PlayerPedId()) if #(pos - marker) &lt; 2.0 then DrawMarker(...) end end\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>It runs sixty times a second forever, calculating a distance to something the player is usually nowhere near. Ten resources doing this is ten times the cost, and none of them individually look bad.\u003C\u002Fp>\u003Cp>The fix is to make the wait time depend on how close the player is to caring:\u003C\u002Fp>\u003Cp>\u003Ccode>while true do local sleep = 1000 local pos = GetEntityCoords(PlayerPedId()) local dist = #(pos - marker) if dist &lt; 20.0 then sleep = 0 DrawMarker(...) end Wait(sleep) end\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>Far away, the loop runs once a second. Close up, it runs every frame. The player sees no difference and the idle cost drops by orders of magnitude. This single change accounts for most of the gap between a script that idles at 0.00ms and one that does not.\u003C\u002Fp>\u003Ch3>Prefer events and zones to polling\u003C\u002Fh3>\u003Cp>Better than a cheap loop is no loop. Two options replace most polling:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Zone libraries\u003C\u002Fstrong> — \u003Ccode>ox_lib\u003C\u002Fcode> and similar provide zone primitives that fire an event on enter and exit. The engine does the checking, your code runs twice per visit instead of sixty times a second.\u003C\u002Fli>\u003Cli>\u003Cstrong>Server events\u003C\u002Fstrong> — anything driven by another player's action should arrive as an event, not be discovered by checking.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>A resource built on zones and events genuinely does nothing while the player is elsewhere, which is what a 0.00ms idle reading means.\u003C\u002Fp>\u003Ch3>Cache what does not change\u003C\u002Fh3>\u003Cp>\u003Ccode>PlayerPedId()\u003C\u002Fcode> inside a per-frame loop is a native call sixty times a second for a value that changes only when the player respawns. Cache it and refresh on the respawn event. The same applies to framework objects, config lookups and anything derived from them — resolve once at start, reuse.\u003C\u002Fp>\u003Cp>This matters more in FiveM than in ordinary Lua because native calls cross into the game engine. They are not free, and a loop making six of them per frame is doing three hundred and sixty engine crossings a second to learn nothing new.\u003C\u002Fp>\u003Ch3>Draw only what is visible\u003C\u002Fh3>\u003Cp>Markers, 3D text and NUI all cost while they are on screen. Gate them on distance the same way as the loop above, and stop drawing entirely when the player cannot see the thing. A common mistake is drawing text at a distance where it renders as a single unreadable pixel — the cost is identical to drawing it up close.\u003C\u002Fp>\u003Ch3>Keep server work off the tick\u003C\u002Fh3>\u003Cp>On the server side, the equivalent sin is blocking the tick. A synchronous database query inside a frequently-called handler stalls everything for its duration, and the symptom is every player rubber-banding at the same instant. Use the async API, index the columns you filter on, and do not query inside loops.\u003C\u002Fp>\u003Ch3>Measure your own work\u003C\u002Fh3>\u003Cp>\u003Ccode>resmon 1\u003C\u002Fcode> shows your resource alongside everything else. Read it in three states: idle with the player far away, near the feature, and using the feature. A well-written resource reads 0.00ms in the first state, which is the state the player spends almost all their time in.\u003C\u002Fp>\u003Cp>If you are evaluating a script before buying, this is the test worth running on a copy of your server. Idle cost is the honest number — peak cost during use matters far less because it is brief and rare.\u003C\u002Fp>\u003Ch3>Judging a script you did not write\u003C\u002Fh3>\u003Cp>Without source access, ask for the idle figure and verify it in \u003Ccode>resmon\u003C\u002Fcode> yourself. With source access, search for \u003Ccode>Wait(0)\u003C\u002Fcode> and read every result: each one is a per-frame loop, and each should be inside a distance check that raises the wait when the player is far away. That single grep tells you more about a resource's cost than its feature list.\u003C\u002Fp>\u003Cp>Our \u003Ca href=\"\u002Fproduct\u002F7561111\">FiveM Zone Manager\u003C\u002Fa> is a useful example of the zone-based approach — it exists so other resources can stop polling and react to zone events instead.\u003C\u002Fp>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>For the server-wide view — finding which resource is the problem in the first place — read our \u003Ca href=\"\u002Fblog\u002Ffivem-server-optimization\">FiveM server optimization guide\u003C\u002Fa>.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002Fd7acef8c38a40da9ff191737459fb23f6df83312.png",64,"FiveM Resource Optimization — Write 0.00ms Scripts","How to write efficient FiveM resources: replace per-frame loops with events, scale wait times by distance, cache framework calls, and profile with resmon.",[19,20],7561111,"2026-09-02T14:47:38.891+00:00","2026-11-21T05:12:39.48213+00:00",{"slug":189,"title":190,"excerpt":191,"content":192,"image":193,"category":11,"author":12,"status":13,"reviewed":14,"views":194,"metaTitle":195,"metaDescription":196,"frameworks":197,"productId":22,"publishedAt":198,"updatedAt":199,"date":198},"fivem-server-optimization","FiveM Server Optimization: A Practical Guide","Server lag is measured, not guessed. How to read resmon, find the resources actually costing you frames, and fix the three causes behind most FiveM performance complaints.","\u003Cp>\"The server is lagging\" describes a symptom with at least five unrelated causes, and every hour spent optimising the wrong one is wasted. This guide is about measuring first: finding which of client frame rate, server tick, database latency or network is actually the problem, and then fixing that.\u003C\u002Fp>\u003Ch3>Start with resmon\u003C\u002Fh3>\u003Cp>Type \u003Ccode>resmon 1\u003C\u002Fcode> in the client console. You get a live table of every running resource with its CPU time in milliseconds. This is the single most useful diagnostic in FiveM and most server owners never open it.\u003C\u002Fp>\u003Cp>Read it while doing normal things — driving, opening menus, standing in a busy area. A resource that idles at 0.01ms and spikes to 8ms when a menu opens is a different problem from one that sits at 3ms constantly. Note both the idle cost and the peak.\u003C\u002Fp>\u003Cp>Rough thresholds for a single resource: under 0.05ms idle is fine, 0.1–0.5ms is worth a look, above 1ms sustained is a problem. The numbers add up across forty resources, which is how a server with no single bad script still runs badly.\u003C\u002Fp>\u003Ch3>The three causes behind most complaints\u003C\u002Fh3>\u003Cp>\u003Cstrong>Loops that run when nothing is happening.\u003C\u002Fstrong> The classic FiveM performance bug is a \u003Ccode>while true do ... Wait(0) end\u003C\u002Fcode> loop that checks something every frame regardless of whether it could have changed. A distance check to a marker the player is nowhere near costs the same as one they are standing on. Well-written resources scale their wait time with distance or use events instead of polling.\u003C\u002Fp>\u003Cp>\u003Cstrong>Database queries in the wrong place.\u003C\u002Fstrong> A query inside a loop, or one that runs on every player join without an index, blocks the server tick while it waits. Symptoms are server-side: everyone rubber-bands at the same moment rather than one player having low FPS. Check that the columns you query on are indexed.\u003C\u002Fp>\u003Cp>\u003Cstrong>Entity count.\u003C\u002Fstrong> Every spawned vehicle, prop and ped is network-synchronised. Servers that let vehicles accumulate — no garage return on restart, no cleanup of abandoned cars — reach the entity limit and everything degrades at once. This is why the garage impound question in our \u003Ca href=\"\u002Fblog\u002Fhow-to-add-a-garage-to-fivem\">garage guide\u003C\u002Fa> is a performance question as well as a gameplay one.\u003C\u002Fp>\u003Ch3>Separate client lag from server lag\u003C\u002Fh3>\u003Cp>They feel similar to players and need opposite fixes. Ask two questions: does it affect everyone at the same moment, and does it show in \u003Ccode>resmon\u003C\u002Fcode>?\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>One player, high resmon numbers\u003C\u002Fstrong> — a client-side resource or that player's hardware.\u003C\u002Fli>\u003Cli>\u003Cstrong>Everyone at once, resmon normal\u003C\u002Fstrong> — the server tick, usually database or entity count.\u003C\u002Fli>\u003Cli>\u003Cstrong>Everyone at once, one resource spiking in resmon\u003C\u002Fstrong> — that resource, and you have already found it.\u003C\u002Fli>\u003Cli>\u003Cstrong>Everyone, intermittent, no pattern\u003C\u002Fstrong> — network or host. Check your provider before your code.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Things worth doing before you profile\u003C\u002Fh3>\u003Cul>\u003Cli>\u003Cstrong>Run a recommended artifact, not the newest.\u003C\u002Fstrong> Latest builds are development builds and performance regressions do ship in them.\u003C\u002Fli>\u003Cli>\u003Cstrong>Set \u003Ccode>sv_maxclients\u003C\u002Fcode> honestly.\u003C\u002Fstrong> It is a ceiling, not capacity. A slot count your hardware cannot serve produces a full server that plays badly.\u003C\u002Fli>\u003Cli>\u003Cstrong>Remove what you do not use.\u003C\u002Fstrong> Every started resource costs something. Servers accumulate test resources that never got removed.\u003C\u002Fli>\u003Cli>\u003Cstrong>Check \u003Ccode>OneSync\u003C\u002Fcode> is on the right mode\u003C\u002Fstrong> for your player count. The infinity mode exists for large servers and has costs on small ones.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Buying with performance in mind\u003C\u002Fh3>\u003Cp>Idle cost is the number to ask about before you buy, because it is paid constantly by every player. A script that idles at 1ms costs more over an evening than one that spikes to 5ms when a menu opens twice. Every ATY script is written to idle at 0.00ms, which is measurable in \u003Ccode>resmon\u003C\u002Fcode> on your own server rather than something you have to take on faith.\u003C\u002Fp>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>For the resource-side view of the same problem — what makes an individual script cheap or expensive — read our \u003Ca href=\"\u002Fblog\u002Ffivem-resource-optimization\">FiveM resource optimization guide\u003C\u002Fa>. If you are still assembling the server, the \u003Ca href=\"\u002Fblog\u002Fhow-to-make-a-fivem-server\">setup guide\u003C\u002Fa> covers the configuration decisions that affect performance later.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002F091804c4d87e6936f4c71acc2dc6aa01cf454e68.jpg",49,"FiveM Server Optimization — Find and Fix Real Lag","Optimize a FiveM server with measurement, not guesswork: reading resmon, hitch detection, database bottlenecks, entity limits and artifact choice.",[19,20],"2026-09-02T14:42:38.891+00:00","2026-11-18T05:05:00.860452+00:00",{"slug":201,"title":202,"excerpt":203,"content":204,"image":205,"category":11,"author":12,"status":13,"reviewed":14,"views":206,"metaTitle":207,"metaDescription":208,"frameworks":209,"productId":210,"publishedAt":211,"updatedAt":212,"date":211},"how-to-add-jobs-to-fivem","How to Add Jobs to FiveM","A job is a definition, a grade ladder, a way to get hired and something to actually do. Most servers get the first three right and the fourth wrong.","\u003Cp>Jobs are how a roleplay server gives players a reason to log in tomorrow. Mechanically a job is four things: a definition, a grade ladder, a way to be hired, and work to do. The first three are configuration. The fourth is design, and it is the one that decides whether the job is used.\u003C\u002Fp>\u003Ch3>Defining the job\u003C\u002Fh3>\u003Cp>In QBCore, jobs live in \u003Ccode>qb-core\u002Fshared\u002Fjobs.lua\u003C\u002Fcode>, each with a label, a set of grades, and a payment value per grade. In ESX they live in the \u003Ccode>jobs\u003C\u002Fcode> and \u003Ccode>job_grades\u003C\u002Fcode> database tables with the same shape. Either way you are declaring a name, the grades within it, and what each grade earns.\u003C\u002Fp>\u003Cp>The job name is the string every script will check against, it is case-sensitive, and changing it later means finding every reference. Pick it once and keep it lowercase and unpunctuated — \u003Ccode>mechanic\u003C\u002Fcode>, not \u003Ccode>Mechanic\u003C\u002Fcode> or \u003Ccode>car-mechanic\u003C\u002Fcode>.\u003C\u002Fp>\u003Ch3>Grades and gating\u003C\u002Fh3>\u003Cp>Grades exist so a job can have progression and so features can be restricted. A boss menu at grade 4, a vehicle spawner at grade 2, basic duties at grade 0.\u003C\u002Fp>\u003Cp>The failure mode here is silent. A player on grade 0 who opens a grade-2 menu gets nothing — no message, no error — which is indistinguishable from a broken script. When you add a gate, add a rejection message with it. \"You do not have permission\" turns a support ticket into a player who understands the system.\u003C\u002Fp>\u003Ch3>Getting hired\u003C\u002Fh3>\u003Cp>Three common approaches, in increasing order of effort:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>A boss menu\u003C\u002Fstrong> — an existing employee at a high grade hires and promotes. Best for player-driven servers, but requires someone to be online.\u003C\u002Fli>\u003Cli>\u003Cstrong>A job centre\u003C\u002Fstrong> — an NPC or terminal where players take entry-level jobs freely. Best for onboarding, and the fastest way to get a new player doing something.\u003C\u002Fli>\u003Cli>\u003Cstrong>Admin-assigned\u003C\u002Fstrong> — staff set jobs by command. Fine for whitelisted roles like police and EMS, unworkable as the only method.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Most servers end up with all three: a job centre for open work, boss menus for the businesses, admin assignment for whitelisted departments.\u003C\u002Fp>\u003Ch3>Designing work players return to\u003C\u002Fh3>\u003Cp>This is the part configuration cannot solve. A job that is one repeated action with a fixed payout is finished the first evening. The jobs that hold players share a few properties:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>A loop with variation\u003C\u002Fstrong> — different locations, different cargo, different routes, so the tenth run is not identical to the first.\u003C\u002Fli>\u003Cli>\u003Cstrong>Progression that changes what you do\u003C\u002Fstrong> — not just a higher payout but access to a different vehicle, a longer route, a harder delivery.\u003C\u002Fli>\u003Cli>\u003Cstrong>A reason to be near other players\u003C\u002Fstrong> — a job that puts two players in the same place produces roleplay; a job in an empty field produces a spreadsheet.\u003C\u002Fli>\u003Cli>\u003Cstrong>Honest pay\u003C\u002Fstrong> — measured against your economy, not copied from another server's config. A job that pays double everything else becomes the only job.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Payment and the server economy\u003C\u002Fh3>\u003Cp>Every job you add is a money faucet. If payouts are set by copying whatever a script shipped with, your economy drifts within weeks and the fix — cutting pay after players have earned it — is far more unpopular than setting it low at the start. Decide what an hour of play should be worth on your server, then set every job against that number.\u003C\u002Fp>\u003Ch3>Adding several jobs at once\u003C\u002Fh3>\u003Cp>Building each job from scratch is the most reliable way to end up with three jobs and a to-do list. A job pack gives you a set that already share a code base, a config style and a balance baseline, which means adding the fourth job is a config edit rather than a project. Our \u003Ca href=\"\u002Fproduct\u002F7167623\">FiveM Advanced Job Pack\u003C\u002Fa> is built that way and detects QBCore and ESX automatically; individual jobs such as the \u003Ca href=\"\u002Fproduct\u002F7443118\">Fishing &amp; Diving Job\u003C\u002Fa>, \u003Ca href=\"\u002Fproduct\u002F6735142\">Bus Job\u003C\u002Fa> and \u003Ca href=\"\u002Fproduct\u002F6893743\">House Cleaner Job\u003C\u002Fa> follow the same structure if you would rather add them one at a time.\u003C\u002Fp>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>Jobs pair naturally with vehicles and a place to store them — see \u003Ca href=\"\u002Fblog\u002Fhow-to-add-a-garage-to-fivem\">how to add a garage to FiveM\u003C\u002Fa>. If jobs are the first thing you are adding to a fresh server, the \u003Ca href=\"\u002Fblog\u002Fhow-to-make-a-fivem-server\">setup guide\u003C\u002Fa> covers what should already be running.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002F4114e788b3d88666607e89b16d80fa262c7a4eb9.png",46,"How to Add Jobs to Your FiveM Server (QBCore & ESX)","Add jobs to FiveM: define the job and grades in QBCore or ESX, gate features by grade, handle hiring and payment, and design work players return to.",[19,20],7167623,"2026-09-02T14:37:38.891+00:00","2026-11-20T18:25:32.920291+00:00",{"slug":214,"title":215,"excerpt":216,"content":217,"image":218,"category":11,"author":12,"status":13,"reviewed":14,"views":219,"metaTitle":220,"metaDescription":221,"frameworks":222,"productId":223,"publishedAt":224,"updatedAt":225,"date":224},"how-to-add-a-garage-to-fivem","How to Add a Garage to FiveM","A garage is a vehicle table, a storage zone and a retrieval zone. Here is what each part does, what breaks when they disagree, and how to place garages without editing code.","\u003Cp>A garage looks simple from the player's side — drive in, the car disappears, come back later and it is there. Underneath it is three separate systems that have to agree: a database record of who owns which vehicle, a zone that stores it, and a zone that gives it back. Nearly every garage bug is those three disagreeing.\u003C\u002Fp>\u003Ch3>How ownership actually works\u003C\u002Fh3>\u003Cp>Owned vehicles live in a database table keyed on the player's identifier and the vehicle's plate. The row holds the plate, the owner, the vehicle's properties as a serialised blob, and a state flag — in the garage, out, or impounded.\u003C\u002Fp>\u003Cp>The properties blob is the part worth understanding. It holds the modifications, colour, and damage of the vehicle at the moment it was stored. When the garage returns the car, it spawns a fresh vehicle and applies that blob. This is why a car sometimes comes back without its tuning: the blob was written before the modification was applied, or the script that applied it did not write back.\u003C\u002Fp>\u003Cp>The plate is the join key, and plates must therefore be unique. A server that generates duplicate plates will eventually hand one player another player's car, and the bug will look random because it only fires when both cars are stored at once.\u003C\u002Fp>\u003Ch3>Storage and retrieval zones\u003C\u002Fh3>\u003Cp>A garage needs a place to put cars in and a place to take them out. Most scripts use the same zone for both, but they are separate operations and can fail independently — a garage that stores but does not return usually has a working zone and a broken query.\u003C\u002Fp>\u003Cp>Place zones somewhere with room to spawn a vehicle without clipping into geometry. A retrieval point pressed against a wall will spawn the car inside it, and the physics response ranges from the car popping out to the player being launched. Test each garage location by retrieving the largest vehicle your server allows, not the smallest.\u003C\u002Fp>\u003Ch3>Impound and the missing-car problem\u003C\u002Fh3>\u003Cp>Every garage needs an answer for a vehicle that is out of the garage when the server restarts. Without one, the row says \"out\", the world says nothing exists, and the player has lost a car they paid for.\u003C\u002Fp>\u003Cp>The two common answers are an impound lot the player pays to recover from, and a restart hook that flips every out-of-garage vehicle back to stored. The impound is better roleplay and the auto-return produces fewer support tickets. Whichever you choose, make sure it exists before launch — this is the single most common cause of \"the garage ate my car\" reports.\u003C\u002Fp>\u003Ch3>Placing garages without editing code\u003C\u002Fh3>\u003Cp>Older garage scripts hard-code coordinates in a config file, which means every new garage is a code edit, a restart, and a trip in game to check you got the numbers right. Newer scripts let you place zones from inside the game and write the config for you.\u003C\u002Fp>\u003Cp>Our \u003Ca href=\"\u002Fproduct\u002F7557947\">FiveM Garage Script &amp; Ownable Garages\u003C\u002Fa> takes that approach: garages are placed in game, and it adds player-owned garages on top — a player can own the building rather than just use a public one. It detects QBCore and ESX automatically and runs at 0.00ms idle, which matters for a resource that is running a zone check for every player at all times.\u003C\u002Fp>\u003Ch3>Testing a garage properly\u003C\u002Fh3>\u003Cul>\u003Cli>Store a modified vehicle, restart the server, retrieve it. The modifications must survive.\u003C\u002Fli>\u003Cli>Store a vehicle, drive it out, restart while it is out. It must be recoverable.\u003C\u002Fli>\u003Cli>Have two players store cars simultaneously. Neither should receive the other's.\u003C\u002Fli>\u003Cli>Retrieve the largest allowed vehicle at every garage. None should spawn inside geometry.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Four tests, ten minutes, and they cover the failures that otherwise arrive as tickets from players who lost progress.\u003C\u002Fp>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>Garages usually go in alongside a vehicle shop and a spawn system. If you are still building the base, the \u003Ca href=\"\u002Fblog\u002Fhow-to-make-a-fivem-server\">server setup guide\u003C\u002Fa> covers the order to add things in, and \u003Ca href=\"\u002Fblog\u002Fhow-to-add-jobs-to-fivem\">adding jobs\u003C\u002Fa> is normally the next step after vehicles.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002F1bfac3321d6c1784236cb401f50a5c210197a06f.png",57,"How to Add a Garage to Your FiveM Server","Add a working garage to FiveM: how owned vehicles are stored, placing garage zones, impound handling, and the common causes of vehicles vanishing.",[19,20],7557947,"2026-09-02T14:32:38.891+00:00","2026-11-20T09:29:01.689127+00:00",{"slug":227,"title":228,"excerpt":229,"content":230,"image":231,"category":11,"author":12,"status":13,"reviewed":14,"views":232,"metaTitle":233,"metaDescription":234,"frameworks":235,"productId":236,"publishedAt":237,"updatedAt":238,"date":237},"qbcore-vs-esx","QBCore vs ESX: Which FiveM Framework Should You Use?","Both are free, mature and widely supported. The real differences are how much ships out of the box, where items live, and how much of your server you end up writing yourself.","\u003Cp>QBCore and ESX are the two frameworks nearly every roleplay server is built on. Both are free, both are actively maintained, both have thousands of compatible scripts. Choosing between them is not a question of which is better — it is a question of how much you want handed to you versus how much you want to decide yourself.\u003C\u002Fp>\u003Ch3>What you get out of the box\u003C\u002Fh3>\u003Cp>QBCore ships as a fuller package. A fresh install already has an inventory, a phone-adjacent set of resources, jobs, a garage, policing and medical resources, and a shared item list. You can put players on a server and have recognisable roleplay running the same day.\u003C\u002Fp>\u003Cp>ESX ships leaner. You get the core, player management, a job system and a menu library; most gameplay is something you add. That is more work up front and more freedom later — nothing to strip out before you build your own version.\u003C\u002Fp>\u003Cp>Neither approach is wrong. If your server has a specific identity you already know, ESX's smaller base means less to fight. If you want a working roleplay server quickly and will customise from there, QBCore gets you further before you write anything.\u003C\u002Fp>\u003Ch3>Where items live\u003C\u002Fh3>\u003Cp>This is the difference you feel daily. QBCore keeps items in a shared Lua file (\u003Ccode>qb-core\u002Fshared\u002Fitems.lua\u003C\u002Fcode>), so adding one is an edit and a restart. Classic ESX keeps them in a database table, so adding one is an SQL insert.\u003C\u002Fp>\u003Cp>The Lua file is easier to version-control and diff, which matters once more than one person edits the server. The database table is easier to change without a restart. In practice many servers on both frameworks end up running \u003Ccode>ox_inventory\u003C\u002Fcode>, which moves items into its own data file and makes this difference mostly disappear.\u003C\u002Fp>\u003Ch3>Jobs and grades\u003C\u002Fh3>\u003Cp>Both use a job name plus a numeric grade, and both gate features on that pair. QBCore defines jobs in \u003Ccode>shared\u002Fjobs.lua\u003C\u002Fcode>; ESX defines them in database tables. The mental model is identical, so job logic ports between them with modest effort — this is not the part that makes migration painful.\u003C\u002Fp>\u003Ch3>Script availability\u003C\u002Fh3>\u003Cp>Both ecosystems are large. Most commercial scripts written in the last few years support both frameworks and detect which one is running, so framework choice rarely locks you out of a specific script any more. Free community resources are more likely to target one, and older free resources skew ESX simply because ESX is older.\u003C\u002Fp>\u003Cp>Every ATY script detects both automatically, so a purchase does not commit you either way.\u003C\u002Fp>\u003Ch3>The migration question\u003C\u002Fh3>\u003Cp>Switching frameworks after launch is expensive, and the cost is not in the scripts — it is in the data. Player money, inventories, owned vehicles, owned property and job progression all live in framework-shaped tables. Migrating them means writing a conversion for each, testing it against real player data, and accepting that anything you miss is a player's progress you deleted.\u003C\u002Fp>\u003Cp>This is why the choice matters more than the feature comparison suggests. Pick the one whose defaults you are happy to live with, because the realistic time to change your mind is before you have players, not after.\u003C\u002Fp>\u003Ch3>A short answer\u003C\u002Fh3>\u003Cul>\u003Cli>\u003Cstrong>New to running a server\u003C\u002Fstrong> — QBCore. More works immediately, and there is more documentation aimed at people starting out.\u003C\u002Fli>\u003Cli>\u003Cstrong>You have a specific server design in mind\u003C\u002Fstrong> — ESX. Less to remove before you build your own.\u003C\u002Fli>\u003Cli>\u003Cstrong>You already know one\u003C\u002Fstrong> — use that one. Familiarity beats the differences described above.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>Once you have chosen, the install guides cover the framework-specific details: \u003Ca href=\"\u002Fblog\u002Fhow-to-install-qbcore-scripts\">installing QBCore scripts\u003C\u002Fa> and \u003Ca href=\"\u002Fblog\u002Fhow-to-install-esx-scripts\">installing ESX scripts\u003C\u002Fa>. Our \u003Ca href=\"\u002Fproduct\u002F7641214\">FiveM Inventory &amp; In-game Editor\u003C\u002Fa> runs on both, so it is a safe first purchase whichever way you go.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002Fa820353c36c289632c0c21c3776ee894cf1b09e9.webp",47,"QBCore vs ESX — Which FiveM Framework to Choose","A practical comparison of QBCore and ESX for FiveM: what ships out of the box, item storage, job systems, script availability and migration cost.",[19,20],7641214,"2026-09-02T14:27:38.891+00:00","2026-11-19T06:06:17.363995+00:00",{"slug":240,"title":241,"excerpt":242,"content":243,"image":244,"category":11,"author":12,"status":13,"reviewed":14,"views":206,"metaTitle":245,"metaDescription":246,"frameworks":247,"productId":22,"publishedAt":248,"updatedAt":249,"date":248},"how-to-make-a-fivem-server","How to Make a FiveM Server","From a Cfx.re license key to players connecting: server artifacts, database, framework choice and the config lines that decide whether anyone can join.","\u003Cp>A FiveM server is four things stacked on each other: the server binaries, a database, a framework, and a config file that ties them together. You can get all four running in an evening. Keeping them running well is a longer job, and this guide flags the decisions that are expensive to reverse later.\u003C\u002Fp>\u003Ch3>1. Get a Cfx.re license key\u003C\u002Fh3>\u003Cp>Every server needs a key from the Cfx.re Keymaster portal, tied to your forum account. The key goes in \u003Ccode>server.cfg\u003C\u002Fcode> as \u003Ccode>sv_licenseKey\u003C\u002Fcode>. Two things worth knowing before you generate it: a key is bound to an IP or a range depending on the type you pick, and escrow-protected paid scripts are tied to the account that owns them. If you plan to buy assets, buy them on the same Cfx.re account that owns the server key — moving assets between accounts later is not something you can do yourself.\u003C\u002Fp>\u003Ch3>2. Install the server artifacts\u003C\u002Fh3>\u003Cp>Download the server build for your platform from the Cfx.re artifacts server. Take a recommended build rather than the newest one: the latest artifact is a development build, and half the \"my server broke overnight\" reports are people running one. Unzip it into its own folder, separate from your server data — you will replace this folder on every update and you do not want your resources inside it.\u003C\u002Fp>\u003Cp>Your directory layout should end up looking like a \u003Ccode>server\u003C\u002Fcode> folder holding the artifacts and a \u003Ccode>server-data\u003C\u002Fcode> folder holding \u003Ccode>resources\u003C\u002Fcode>, \u003Ccode>server.cfg\u003C\u002Fcode> and your cache. Keeping them apart makes updating the artifacts a delete-and-replace instead of a careful merge.\u003C\u002Fp>\u003Ch3>3. Set up the database\u003C\u002Fh3>\u003Cp>You need MySQL or MariaDB, and MariaDB is the common choice. Install it, create a database for the server, create a user with rights on it, and note the connection string — every framework and most scripts will need it. In \u003Ccode>server.cfg\u003C\u002Fcode> that looks like:\u003C\u002Fp>\u003Cp>\u003Ccode>set mysql_connection_string \"mysql:\u002F\u002Fuser:password@localhost\u002Fservername?charset=utf8mb4\"\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>Use \u003Ccode>utf8mb4\u003C\u002Fcode>. The older \u003Ccode>utf8\u003C\u002Fcode> charset in MySQL cannot store four-byte characters, which means emoji in player names or chat will throw errors that look like database corruption.\u003C\u002Fp>\u003Ch3>4. Choose a framework\u003C\u002Fh3>\u003Cp>QBCore or ESX. This is the decision that is hardest to reverse, because every script you buy, every job you write and every database table you fill assumes one of them. Both are mature, both are free, both have large script ecosystems. Our \u003Ca href=\"\u002Fblog\u002Fqbcore-vs-esx\">QBCore vs ESX comparison\u003C\u002Fa> goes through the practical differences; the short version is that QBCore ships more gameplay out of the box while ESX gives you a leaner base to build on.\u003C\u002Fp>\u003Cp>Whichever you pick, install it as a resource, import its SQL, and start it before everything else. Get a player spawning and saving before you add a single extra script — a framework that half-works produces errors in every resource above it and you will chase them in the wrong place.\u003C\u002Fp>\u003Ch3>5. Write server.cfg\u003C\u002Fh3>\u003Cp>The lines that decide whether anyone can actually join:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ccode>endpoint_add_tcp \"0.0.0.0:30120\"\u003C\u002Fcode> and \u003Ccode>endpoint_add_udp \"0.0.0.0:30120\"\u003C\u002Fcode> — bind on all interfaces, not localhost\u003C\u002Fli>\u003Cli>\u003Ccode>sv_maxclients 48\u003C\u002Fcode> — the ceiling; raising it does not create capacity\u003C\u002Fli>\u003Cli>\u003Ccode>sv_hostname\u003C\u002Fcode>, \u003Ccode>sets sv_projectName\u003C\u002Fcode>, \u003Ccode>sets sv_projectDesc\u003C\u002Fcode> — how you appear in the server list\u003C\u002Fli>\u003Cli>\u003Ccode>sv_licenseKey\u003C\u002Fcode> — the key from step 1\u003C\u002Fli>\u003Cli>\u003Ccode>ensure\u003C\u002Fcode> lines for every resource, in dependency order\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Then open port 30120 on both TCP and UDP in your firewall and, if you are behind one, your router. A server that starts cleanly and appears in no list is almost always a closed UDP port — TCP alone lets the server boot but not be discovered.\u003C\u002Fp>\u003Ch3>6. Add scripts, one at a time\u003C\u002Fh3>\u003Cp>Resist installing twenty resources before the first test. Add one, restart, watch the console, join, confirm it works. A server built one resource at a time takes an evening; a server built all at once and then debugged takes a week, because every error could come from anything.\u003C\u002Fp>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>Once the base runs, the first things most servers add are a spawn and character system, a garage, and jobs. Our guides cover \u003Ca href=\"\u002Fblog\u002Fhow-to-add-a-garage-to-fivem\">adding a garage\u003C\u002Fa> and \u003Ca href=\"\u002Fblog\u002Fhow-to-add-jobs-to-fivem\">adding jobs\u003C\u002Fa>, and the \u003Ca href=\"\u002Fblog\u002Ffivem-server-optimization\">server optimization guide\u003C\u002Fa> covers what to measure before your player count grows.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002F07fb70d7436b7fba6ef359f7515b4a79b28886df.jpg","How to Make a FiveM Server (Complete Setup Guide)","Build a FiveM server from scratch: get a Cfx.re key, install artifacts and MariaDB, pick QBCore or ESX, configure server.cfg and open the right ports.",[19,20],"2026-09-02T14:22:38.891+00:00","2026-11-19T15:19:42.600443+00:00",{"slug":251,"title":252,"excerpt":253,"content":254,"image":255,"category":11,"author":12,"status":13,"reviewed":14,"views":256,"metaTitle":257,"metaDescription":258,"frameworks":259,"productId":260,"publishedAt":261,"updatedAt":262,"date":261},"how-to-install-esx-scripts","How to Install ESX Scripts","ESX installs hinge on three things the readme rarely spells out: which ESX version you run, where items live, and how the shared object is obtained. Here is all three.","\u003Cp>ESX has been through enough versions that \"install an ESX script\" means slightly different things depending on which ESX you run. Most install problems trace back to a version mismatch rather than a mistake in the steps, so this guide starts there.\u003C\u002Fp>\u003Ch3>Know which ESX you are running\u003C\u002Fh3>\u003Cp>Older scripts obtain the shared object by looping on an event:\u003C\u002Fp>\u003Cp>\u003Ccode>TriggerEvent('esx:getSharedObject', function(obj) ESX = obj end)\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>Current ESX exposes it directly:\u003C\u002Fp>\u003Cp>\u003Ccode>ESX = exports['es_extended']:getSharedObject()\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>A script written for one style running on the other produces a nil ESX object and fails on its first call. If you buy a script and it errors immediately with \u003Ccode>attempt to index a nil value (global 'ESX')\u003C\u002Fcode>, check which retrieval method it uses before assuming the script is broken — many scripts include both paths and pick automatically, but older free resources often do not.\u003C\u002Fp>\u003Ch3>Start order\u003C\u002Fh3>\u003Cp>As with any framework, \u003Ccode>es_extended\u003C\u002Fcode> must start before anything that uses it, and the database layer must start before \u003Ccode>es_extended\u003C\u002Fcode>. A working order looks like:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ccode>ensure oxmysql\u003C\u002Fcode>\u003C\u002Fli>\u003Cli>\u003Ccode>ensure es_extended\u003C\u002Fcode>\u003C\u002Fli>\u003Cli>\u003Ccode>ensure ox_lib\u003C\u002Fcode> — if your scripts use it\u003C\u002Fli>\u003Cli>\u003Ccode>ensure esx_menu_default\u003C\u002Fcode> and the other ESX menu resources, if the script expects them\u003C\u002Fli>\u003Cli>…then gameplay resources\u003C\u002Fli>\u003C\u002Ful>\u003Cp>ESX menus are a real dependency, not decoration. A script that opens an \u003Ccode>ESX.UI.Menu\u003C\u002Fcode> will do nothing at all if the menu resource is missing, with no error in the console — the most confusing failure mode in ESX.\u003C\u002Fp>\u003Ch3>Items live in the database, not a Lua file\u003C\u002Fh3>\u003Cp>This is the biggest practical difference from QBCore. Classic ESX stores items in the \u003Ccode>items\u003C\u002Fcode> table in your database, so adding an item means an SQL insert rather than editing a shared file:\u003C\u002Fp>\u003Cp>\u003Ccode>INSERT INTO items (name, label, weight, rare, can_remove) VALUES ('lockpick', 'Lockpick', 1, 0, 1);\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>If you run \u003Ccode>ox_inventory\u003C\u002Fcode> on top of ESX — an increasingly common setup — items move back into \u003Ccode>ox_inventory\u002Fdata\u002Fitems.lua\u003C\u002Fcode> and the database table stops being the source of truth. Know which one your server uses, because adding an item to the wrong place produces a script that runs and an item that does not exist.\u003C\u002Fp>\u003Ch3>Jobs and grades\u003C\u002Fh3>\u003Cp>ESX jobs live in the \u003Ccode>jobs\u003C\u002Fcode> and \u003Ccode>job_grades\u003C\u002Fcode> tables. A script gated to a job checks \u003Ccode>xPlayer.job.name\u003C\u002Fcode>, so the job row must exist and the grade must be within the range the script expects. Missing grades are a quiet failure: the job exists, the player has it, and the menu still refuses to open because the grade check falls through.\u003C\u002Fp>\u003Ch3>Import SQL, then restart fully\u003C\u002Fh3>\u003Cp>Run every \u003Ccode>.sql\u003C\u002Fcode> file the script ships before its first start. Then restart the server rather than hot-reloading — ESX caches player data on join, so a script that adds columns to a player table needs players to reconnect before it sees them.\u003C\u002Fp>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>ATY scripts detect ESX and QBCore automatically, so the same download works on either. If you have not committed to a framework yet, read \u003Ca href=\"\u002Fblog\u002Fqbcore-vs-esx\">QBCore vs ESX\u003C\u002Fa> first. The \u003Ca href=\"\u002Fproduct\u002F6396389\">FiveM Vehicle Shop Script\u003C\u002Fa> is a good first install on ESX: it touches the database, jobs and the inventory, so getting it working proves the whole chain.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002F9c94e3188b4b9855f2d25bf266c220d1f9fb8e7c.png",43,"How to Install ESX Scripts on a FiveM Server","Install ESX scripts properly: legacy vs new ESX object retrieval, adding items to the database, registering jobs, and fixing the common nil ESX error.",[20],6396389,"2026-09-02T14:17:38.891+00:00","2026-11-19T13:25:48.845409+00:00",{"slug":264,"title":265,"excerpt":266,"content":267,"image":218,"category":11,"author":12,"status":13,"reviewed":14,"views":268,"metaTitle":269,"metaDescription":270,"frameworks":271,"productId":223,"publishedAt":272,"updatedAt":273,"date":272},"how-to-install-qbcore-scripts","How to Install QBCore Scripts","QBCore scripts need more than an ensure line: shared item entries, job definitions and the right start order. This is the QBCore-specific half of a script install.","\u003Cp>The generic install steps — unzip, \u003Ccode>ensure\u003C\u002Fcode>, restart — get a QBCore script running. What they do not do is register its items, its jobs or its permissions, and that is where most QBCore installs actually stall. This guide covers the QBCore-specific work that the readme usually assumes you already know.\u003C\u002Fp>\u003Ch3>Start order is not optional\u003C\u002Fh3>\u003Cp>QBCore scripts read a shared core object at start. If \u003Ccode>qb-core\u003C\u002Fcode> has not started yet, that object does not exist and every script above it fails with \u003Ccode>attempt to index a nil value (global 'QBCore')\u003C\u002Fcode>. Your \u003Ccode>server.cfg\u003C\u002Fcode> should read roughly in this order:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ccode>ensure oxmysql\u003C\u002Fcode> — the database layer, before anything that queries\u003C\u002Fli>\u003Cli>\u003Ccode>ensure qb-core\u003C\u002Fcode> — the framework itself\u003C\u002Fli>\u003Cli>\u003Ccode>ensure ox_lib\u003C\u002Fcode> — shared UI and callback library, if your scripts use it\u003C\u002Fli>\u003Cli>\u003Ccode>ensure qb-inventory\u003C\u002Fcode> (or \u003Ccode>ox_inventory\u003C\u002Fcode>) — whatever holds items\u003C\u002Fli>\u003Cli>…then every gameplay resource\u003C\u002Fli>\u003C\u002Ful>\u003Cp>The error is misleading because it names the script that failed rather than the framework that was missing, so it sends people to the wrong readme. If a brand new script throws a nil QBCore error, check the order before you check the script.\u003C\u002Fp>\u003Ch3>Adding items to shared\u002Fitems.lua\u003C\u002Fh3>\u003Cp>Anything a script gives, takes or consumes must exist as an item first. In QBCore that means an entry in \u003Ccode>qb-core\u002Fshared\u002Fitems.lua\u003C\u002Fcode>, and the key you write there is the name the script will reference in its own config:\u003C\u002Fp>\u003Cp>\u003Ccode>['lockpick'] = { name = 'lockpick', label = 'Lockpick', weight = 100, type = 'item', image = 'lockpick.png', unique = false, useable = true, shouldClose = true, description = 'Opens things that are not yours' },\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>Three fields cause most of the trouble. \u003Ccode>image\u003C\u002Fcode> must match a real file in your inventory resource's image folder, or the item shows as a blank square. \u003Ccode>useable\u003C\u002Fcode> must be true for anything the player is meant to click. And \u003Ccode>unique\u003C\u002Fcode> controls whether the item stacks — get this wrong on a crafting item and players will either lose stacks or duplicate them.\u003C\u002Fp>\u003Cp>If you run \u003Ccode>ox_inventory\u003C\u002Fcode> instead of \u003Ccode>qb-inventory\u003C\u002Fcode>, items live in \u003Ccode>ox_inventory\u002Fdata\u002Fitems.lua\u003C\u002Fcode> in a different format. Scripts that support both will say which one they expect; putting the item in the wrong file produces a script that runs but never finds its own items.\u003C\u002Fp>\u003Ch3>Registering jobs and grades\u003C\u002Fh3>\u003Cp>Job-based scripts check \u003Ccode>PlayerData.job.name\u003C\u002Fcode> against a string in their config. That job has to exist in \u003Ccode>qb-core\u002Fshared\u002Fjobs.lua\u003C\u002Fcode> with matching grades, or the check silently fails and nothing opens. The job name is case-sensitive and must match exactly — \u003Ccode>mechanic\u003C\u002Fcode> and \u003Ccode>Mechanic\u003C\u002Fcode> are different jobs to QBCore.\u003C\u002Fp>\u003Cp>Grades matter for permission gates. A script that restricts a menu to grade 3 and above will do nothing for a player on grade 0, which looks identical to a broken script from the player's side. When testing, set yourself to the highest grade first, confirm the feature works, then test the gate.\u003C\u002Fp>\u003Ch3>Import the SQL before the first start\u003C\u002Fh3>\u003Cp>QBCore scripts that persist anything — owned vehicles, owned property, job data — ship a \u003Ccode>.sql\u003C\u002Fcode> file. Import it into your database before starting the resource. Starting first produces database errors on the first player interaction rather than at boot, which makes them look like gameplay bugs instead of a missed setup step.\u003C\u002Fp>\u003Ch3>Checking that it actually loaded\u003C\u002Fh3>\u003Cp>After a restart, run \u003Ccode>ensure\u003C\u002Fcode> again from the console on just that resource and watch the output. A resource that starts clean prints nothing after its name. Any Lua error at start means the script never registered its exports, and every later interaction will fail regardless of your config.\u003C\u002Fp>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>Every ATY script detects QBCore automatically and ships with a QBCore section in its docs. If you are weighing QBCore against ESX before committing, our \u003Ca href=\"\u002Fblog\u002Fqbcore-vs-esx\">QBCore vs ESX comparison\u003C\u002Fa> covers what actually differs in practice. For a worked example, the \u003Ca href=\"\u002Fproduct\u002F7557947\">FiveM Garage Script &amp; Ownable Garages\u003C\u002Fa> install is documented end to end.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>",36,"How to Install QBCore Scripts on a FiveM Server","Install QBCore scripts correctly: qb-core start order, adding items to shared\u002Fitems.lua, registering jobs in shared\u002Fjobs.lua, and fixing nil QBCore errors.",[19],"2026-09-02T14:12:38.891+00:00","2026-11-18T03:38:49.665908+00:00",{"slug":275,"title":276,"excerpt":277,"content":278,"image":231,"category":11,"author":12,"status":13,"reviewed":14,"views":279,"metaTitle":280,"metaDescription":281,"frameworks":282,"productId":22,"publishedAt":283,"updatedAt":284,"date":283},"how-to-install-a-fivem-script","How to Install a FiveM Script","Every FiveM script installs the same way: drop the folder in resources, start it in server.cfg, restart. Here is the full process, plus the four mistakes that cause most \"resource not found\" errors.","\u003Cp>Installing a FiveM script is the same four steps whether the resource costs nothing or costs fifty euros: put the folder somewhere the server reads, tell the server to start it, satisfy its dependencies, restart. Almost every \"it doesn't work\" message comes from getting one of those four slightly wrong, so this guide walks each one and names the specific error you will see when it fails.\u003C\u002Fp>\u003Ch3>1. Put the folder in your resources directory\u003C\u002Fh3>\u003Cp>Unzip the download and drop the folder into your server's \u003Ccode>resources\u003C\u002Fcode> directory. Most server owners keep third-party scripts in a subfolder like \u003Ccode>resources\u002F[aty]\u003C\u002Fcode> or \u003Ccode>resources\u002F[scripts]\u003C\u002Fcode> — FiveM reads any folder wrapped in square brackets recursively, so nesting is free organisation.\u003C\u002Fp>\u003Cp>Two rules matter here. The folder name must match the name you will use in \u003Ccode>server.cfg\u003C\u002Fcode>, and it must not contain spaces or capital letters that you then type differently. If your download unzipped into \u003Ccode>aty_garage-main\u003C\u002Fcode> or \u003Ccode>aty_garage (1)\u003C\u002Fcode>, rename it to \u003Ccode>aty_garage\u003C\u002Fcode> now — the trailing \u003Ccode>-main\u003C\u002Fcode> from a zip download is the single most common cause of a resource that \"won't start\".\u003C\u002Fp>\u003Ch3>2. Start it in server.cfg\u003C\u002Fh3>\u003Cp>Open \u003Ccode>server.cfg\u003C\u002Fcode> and add a line for the resource:\u003C\u002Fp>\u003Cp>\u003Ccode>ensure aty_garage\u003C\u002Fcode>\u003C\u002Fp>\u003Cp>Use \u003Ccode>ensure\u003C\u002Fcode> rather than \u003Ccode>start\u003C\u002Fcode>. They look interchangeable but they are not: \u003Ccode>ensure\u003C\u002Fcode> stops the resource first if it is already running, which means a restart actually reloads your changes instead of silently keeping the old version in memory.\u003C\u002Fp>\u003Cp>\u003Cstrong>Order matters.\u003C\u002Fstrong> A resource that depends on another must be listed after it. Your framework (\u003Ccode>qb-core\u003C\u002Fcode> or \u003Ccode>es_extended\u003C\u002Fcode>) and shared libraries (\u003Ccode>ox_lib\u003C\u002Fcode>, \u003Ccode>oxmysql\u003C\u002Fcode>) belong near the top of the file, above every script that uses them. If you put a job script above \u003Ccode>qb-core\u003C\u002Fcode>, it will start, fail to find the framework object, and throw an error that looks like a bug in the script.\u003C\u002Fp>\u003Ch3>3. Install the dependencies\u003C\u002Fh3>\u003Cp>Read the resource's documentation before the first start, not after the first error. Most modern FiveM scripts expect some combination of:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>A framework\u003C\u002Fstrong> — QBCore (\u003Ccode>qb-core\u003C\u002Fcode>) or ESX (\u003Ccode>es_extended\u003C\u002Fcode>). Many scripts support both and detect which one you run automatically.\u003C\u002Fli>\u003Cli>\u003Cstrong>A database connector\u003C\u002Fstrong> — \u003Ccode>oxmysql\u003C\u002Fcode> is the current standard. If the script ships SQL files, run them against your database before starting the server.\u003C\u002Fli>\u003Cli>\u003Cstrong>A UI\u002Futility library\u003C\u002Fstrong> — \u003Ccode>ox_lib\u003C\u002Fcode> for menus, notifications and callbacks is the most common.\u003C\u002Fli>\u003Cli>\u003Cstrong>An inventory\u003C\u002Fstrong> — anything that gives or takes items needs one, and the script's config usually has a setting to pick which.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Missing dependencies produce errors that name the missing thing, so read the console output rather than guessing. A line containing \u003Ccode>attempt to index a nil value\u003C\u002Fcode> almost always means a dependency did not start.\u003C\u002Fp>\u003Ch3>4. Restart and read the console\u003C\u002Fh3>\u003Cp>Restart the server fully for the first install — a live \u003Ccode>refresh; ensure aty_garage\u003C\u002Fcode> works for updates but can miss newly added dependencies. Then watch the console as the server boots. A clean start prints the resource name with no error block after it.\u003C\u002Fp>\u003Cp>If you bought an escrow-protected script, it also needs to be linked to your server's Cfx.re key. Escrow resources are tied to the license in your Keymaster account, so the server must run on a key that owns the asset; running it on a different key produces a decryption error rather than a Lua error.\u003C\u002Fp>\u003Ch3>The four errors you will actually hit\u003C\u002Fh3>\u003Cul>\u003Cli>\u003Cstrong>\"Could not find resource\"\u003C\u002Fstrong> — the folder name and the \u003Ccode>ensure\u003C\u002Fcode> line disagree, usually because of a \u003Ccode>-main\u003C\u002Fcode> suffix or a typo.\u003C\u002Fli>\u003Cli>\u003Cstrong>\"attempt to index a nil value (global 'QBCore')\"\u003C\u002Fstrong> — the framework starts after the script. Move it up in \u003Ccode>server.cfg\u003C\u002Fcode>.\u003C\u002Fli>\u003Cli>\u003Cstrong>Nothing happens, no error\u003C\u002Fstrong> — the resource started but its trigger is not bound. Check the config for the command, key or zone that opens it.\u003C\u002Fli>\u003Cli>\u003Cstrong>Database errors on first use\u003C\u002Fstrong> — the SQL file was never imported. Run it, then restart.\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>Where to go next\u003C\u002Fh3>\u003Cp>If you are setting up a server from scratch rather than adding one script to an existing one, start with our guide on \u003Ca href=\"\u002Fblog\u002Fhow-to-make-a-fivem-server\">how to make a FiveM server\u003C\u002Fa>. If you already know your framework, the framework-specific walkthroughs cover the details that differ: \u003Ca href=\"\u002Fblog\u002Fhow-to-install-qbcore-scripts\">QBCore\u003C\u002Fa> and \u003Ca href=\"\u002Fblog\u002Fhow-to-install-esx-scripts\">ESX\u003C\u002Fa>.\u003C\u002Fp>\u003Cp>Every ATY script ships with its own setup page in the \u003Ca href=\"\u002Fdocs\">documentation\u003C\u002Fa>, and delivery is automatic through Cfx.re Keymaster the moment payment clears. Browse the full catalogue in the \u003Ca href=\"\u002Fstore\">store\u003C\u002Fa>, or open a ticket in our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FdvPMYsRFNx\" rel=\"noopener\">Discord\u003C\u002Fa> if you get stuck — you talk to the developers who wrote the code.\u003C\u002Fp>",52,"How to Install a FiveM Script (Step-by-Step Guide)","Install any FiveM script in four steps: unzip into resources, add ensure to server.cfg, install dependencies, restart. Includes fixes for the most common errors.",[19,20],"2026-09-02T14:07:38.891+00:00","2026-11-21T09:37:49.197729+00:00",{"slug":286,"title":287,"excerpt":288,"content":289,"image":290,"category":31,"author":12,"status":13,"reviewed":83,"views":291,"metaTitle":22,"metaDescription":22,"frameworks":292,"productId":112,"publishedAt":293,"updatedAt":294,"date":293},"fivem-map-editor-and-prop-placer","FiveM Map Editor & Prop Placer — Feature Breakdown","ATY Map Editor is a professional, advanced, and feature-rich prop and map editor resource that redefines mapping and world customization on FiveM. Built…","\u003Cp>ATY Map Editor is a professional, advanced, and feature-rich prop and map editor resource that redefines mapping and world customization on FiveM. Built with a stunning Vue.js 3 interface and a robust server-side synchronization architecture, it provides developers and players with state-of-the-art tools to spawn props, customize lights, delete world objects, and batch-create structures in real-time, ensuring a premium roleplay feel and seamless map development for any community.\u003C\u002Fp>\u003Ch3>Key Features\u003C\u002Fh3>\u003Cp>The core capabilities of FiveM Map Editor &amp; Prop Placer:\u003C\u002Fp>\u003Cul>\u003Cli>Advanced Gizmo System (Translation, Rotation &amp; Scaling)\u003C\u002Fli>\u003Cli>Dual Space Coordinate Transformations (World &amp; Local Axis)\u003C\u002Fli>\u003Cli>Cognitive Smart Search Recommendation Engine (Conceptual keyword matching)\u003C\u002Fli>\u003Cli>Interactive 3D World Raycast &amp; Ground Alignment\u003C\u002Fli>\u003Cli>Dynamic Custom Category and Favorites Management System\u003C\u002Fli>\u003Cli>Searchable Prop Model Replacement &amp; Customization Tool\u003C\u002Fli>\u003Cli>Professional Light Source Editor (Spot, Point, Shadow Controls, Intensity &amp; Colors)\u003C\u002Fli>\u003Cli>High-Precision Area Fill (Mass Spawns with Grid or Random Distribution)\u003C\u002Fli>\u003Cli>Area Delete Tool for Instant Rectangular Bounds Cleanup\u003C\u002Fli>\u003Cli>World Eraser Tool to Permanently Suppress GTA V Default Objects\u003C\u002Fli>\u003C\u002Ful>\u003Cp>That is 10 of 15 documented features — the full list is on the product page.\u003C\u002Fp>\u003Ch3>Framework Compatibility\u003C\u002Fh3>\u003Cp>FiveM Map Editor &amp; Prop Placer works with Standalone, ox_lib. Configuration lives in a single config file, so adapting it to a custom framework setup does not require touching the core logic.\u003C\u002Fp>\u003Ch3>Performance & Support\u003C\u002Fh3>\u003Cp>FiveM Map Editor &amp; Prop Placer is optimised to stay at 0.00ms when idle, which matters once you are running dozens of resources side by side. Lifetime updates and Discord support are included.\u003C\u002Fp>\u003Cp>\u003Ca href=\"\u002Fproduct\u002F7471609\">View FiveM Map Editor &amp; Prop Placer in the store →\u003C\u002Fa>\u003C\u002Fp>","https:\u002F\u002Fdunb17ur4ymx4.cloudfront.net\u002Fpackages\u002Fimages\u002Fa501d3021eb73bf963ef3647a0c2545e8da2f096.png",105,[],"2026-07-11T00:00:00+00:00","2026-11-19T17:30:58.312126+00:00","db",23]