The China edition of TRAE combines an AI development workspace with task planning, code changes and previews. Its current free route uses recurring credits rather than unlimited model access.
TRAE CN is useful when you want an AI assistant to work through a development task inside a project rather than only provide a code snippet. The current product family includes TraeCode for development and TraeWork for office work. This guide concentrates on the coding workflow and the China edition’s published credit rules.
The strongest first task is specific enough to review: fix a form validation problem, add one screen or explain a failing build. A broad request to rebuild everything can create a large diff without making the intended outcome clear. Keep a short description of the user-visible behavior, the files likely involved and what counts as a successful result.
A focused interface improvement
An unfamiliar project walkthrough
A contained bug fix
A prototype with an explicit boundary
Official product image. Click to inspect the details.
Choose the right edition and client
Prepare a reversible working state
Review the plan and execution permissions
Inspect code and the visible outcome
Example prompt or task: Improve this existing form’s empty and error states. Keep the current visual style and API contract, explain the files you will change, then show the diff and verify one successful submission and one invalid input.
This is an example project brief. It does not authorize an agent to deploy, purchase services or access unrelated files.
Monthly credits expire in the current calendar month. Registration bonuses and paid plans have separate rules.
The current China billing guide gives free users 500 credits for monthly login, valid in that calendar month, and 150 credits for daily check-in with a 31-day validity. The one-time registration or upgrade reward is separate from these continuing routes.
Built-in model calls consume TRAE credits; custom model calls do not consume those credits but can incur provider charges. Paid plans add capacity and benefits. A credit is not a fixed number of completed features, so estimate needs using the actual model, context and task instead of translating the monthly allowance into a guaranteed project count.
No. Its current free model access is credit-based.
No. Monthly login credits expire during the current calendar month.
No. They are a one-time benefit, separate from recurring credits.
The guide says no, but the external provider may charge.
Do not assume that; this page covers the China edition.
No. Choose execution permissions appropriate to the actual task.
Inspect the changed behavior and relevant verification output.
Explore platforms, inputs and outputs, licensing, and access requirements.
TRAE’s product names and billing have changed, so older descriptions of an entirely unlimited China edition should not be used as current guidance. The global edition can have different accounts, models and plans; the numbers on this page apply to the China documentation reviewed on the stated date.
SOLO accepts task descriptions and supporting material and brings execution and preview into one workspace. Good results still depend on a runnable project, correct environment and clear requirements. If a preview fails, first distinguish an application error from missing local dependencies or credentials before requesting another broad rewrite.
A development agent can read project context and use connected tools. Review what is sent to the model and which services can receive data, especially when a repository includes private customer information or credentials.
The automatic-run guide says automatic terminal commands run outside the sandbox and skip specified checks. Keep that distinction visible when choosing permissions. MCP connections can act on external systems, so attach only services appropriate for the task and retain a clear review boundary for consequential actions.
Reviewed October 3, 2026. Product facts come from the official sources below. Suggested projects, prompts and review methods are FindGoodAI editorial guidance, not measured performance results.
Loading experiences…
Saved tools are private. Approved comments are public; edits return to moderation. Comment counts include approved comments and replies. Each person has one rating. New ratings are automatically approved; withdrawn, unapproved or invalid ratings do not count. Share real experiences and avoid spam, private information or personal attacks.
For moderation appeals or data requests: support@findgoodai.com
Bring a real task and see how it fits the way you work.
User experiences