FiveM development. Step by step.
Larger tasks need a clear plan.
Agent Mode connects your brief with planning, files and a final review. It suits related changes where several parts of a FiveM resource need to stay coordinated.
What you can do
Define a result you can verify: what should change, what must stay intact and how will you recognise a useful outcome?
Open Agent ModePlan → Build → Review
Set the scope of the task
Describe affected resources, framework versions and existing interfaces. Name areas that must remain intact, such as event names or current database tables. A focused brief is easier to plan and review than “optimise my entire server”.
Follow the plan and file changes
Use workspace context for files relevant to the task. Follow the plan and check that its steps address your goal. For changes across several files, make sure client, server and NUI still expect compatible data formats and events.
Measure the result against the requirements
Review the modified files against your agreed requirements. A completed agent task is not automatic production approval. Check available tests and run the relevant flows on your FiveM test server before distributing the updated resource.
For your next task
An example Agent Mode brief
Extend the existing garage with server-checked access to job vehicles. Create a plan first. Preserve the private vehicle menu and current events. Change client, server and NUI only where this feature requires it, and explain the tests needed.
Check before moving on
- Define the goal, scope and interfaces to preserve
- Include only relevant project files
- Review the plan and changes regularly
- Test integration cases before release
Scope and limitations
Availability and usage depend on enabled features and your account. AI requests consume credits. The agent does not replace product decisions or testing under actual server conditions.
Current plans and conditions →Common questions
When is Agent Mode useful instead of a single chat request?
When a defined goal requires several coordinated file changes. AI chat is often sufficient for a short explanation or an individual question.
Do I still need to review the result?
Yes. Review the changes and test the affected gameplay. The agent can help structure checks, but it does not provide a blanket functionality guarantee.
Your project. Your next step.
Describe your idea or upload an existing resource. Continue working on the source code and test the result on your server.
Open AI chat →