Trae

Trae is ByteDance’s AI-native code editor. Like several of its rivals it is built on VS Code, offering completions, chat-driven edits and an agent mode that can carry out multi-step tasks. It comes in an international edition and a China edition; the China edition is directly accessible, and the UI works well in Chinese.

It is positioned much like Cursor and is fine for day-to-day work. Its biggest draw for developers in China is the low barrier: no network workarounds, a fairly generous free allowance, and good understanding of requirements written in Chinese.

Extension compatibility and understanding of very large repos are still catching up, and the two editions do not offer exactly the same models. Trial it on a real project for a while before deciding to switch a whole team.

Key features

• Code completion: real-time suggestions as you type, including predictions of the next edit based on context.
• Chat edits: describe what you want in the sidebar, reference files or a selection, and review the change as a diff.
• Agent mode: describe a whole task and it breaks it into steps, creates and edits multiple files and runs commands.
• Settings import: on first launch it can import settings and extensions from VS Code or Cursor.
• Chinese UI: menus, prompts and docs are in Chinese, which helps developers less comfortable with English interfaces.

How to use

1. Visit the official site, pick the China or international edition for your region, then download and install it.
2. Sign in and follow the prompts to import your existing editor configuration.
3. Open a project and start by asking it to explain the code structure, to get a feel for how well it understands it.
4. Hand a small feature to agent mode, review the changes, and widen its use gradually.

Best for

• Individual developers in China who want an AI editor without network setup.
• Developers who prefer working in Chinese, both in the UI and in prompts.
• Small projects and prototypes, where agent mode quickly produces a working first version.
• People who want to try an AI editor before paying; the free allowance covers everyday experimenting.

Strengths and limitations

Strengths

• The China edition is directly accessible with no network hurdles.
• Good Chinese UI and good comprehension of Chinese prompts.
• A relatively generous free allowance keeps the entry cost low.

Limitations

• Extension compatibility is not yet as complete as VS Code or Cursor; some extensions may not work.
• Understanding of very large repos and the reliability of multi-file edits are still catching up.
• The two editions differ in models and features, so tutorials online may not match the version you use.
• Teams with strict code-privacy requirements should review its data-handling policies themselves.