Code Bridge
Bring your code to an AI chat. Bring back a change you understand.
Start with a project
Choose the folder you use in VS Code. Code Bridge reads its source files so you can include the right code in your request.
New here? Try a small bug.
A jumping game, a ready-made task, and a sample answer. No AI login needed to practice.
How Code Bridge works
- Prepare a request hereSelect files and describe one change. Copy the request into your teacher-approved AI chat.
- Bring the answer backPaste the AI’s answer here. Read the changes, approve each file, then save or download them.
- Test in VS CodeRun your project and check what happened. Record a checkpoint: a note of your test and result.
Your files stay on this device. You choose what to copy into your AI chat. Code Bridge does not connect to AI or run your project.
Choose your project above, or try the jump bug for a guided practice.
What do you want to change?
Select the files the AI needs, then write one goal and a test.
0 selected files. Nothing is sent automatically.
Request size settings
Includes instructions, the file list, notes, and selected code. Nothing is shortened to fit. Token counts are estimates; AI services count differently.
Your request is ready
Read it below, then copy it into your teacher-approved Google AI chat. The request asks for complete files in the format Code Bridge needs.
Once the AI has answered, bring its response back here.
Review complete files before replacing anything
You need a prepared request to compare the AI’s answer with your original code.
What should the AI’s answer look like?
The request already asks for this format: a filename heading, then a code block with the entire file. If the AI asks a question, answer it in the chat first.
### FILE: src/player.js ```js // Complete file contents here ```
Prepare a request first so the answer can be checked against its original files.
Apply the files you approved
0 files approved
Folder saving is available after opening a folder in a supported browser. Downloads work without folder write access.
ZIP exports have changes/ and backup/ folders. Copy files inside changes/ into the matching paths in your VS Code project. Then reload the folder here. An export does not mean you applied or tested the change.
After applying the changes, run your project in VS Code.
Did the change work?
A checkpoint is your record of what you tested and what happened. Run the project in VS Code using your usual setup or Live Server.
Write a success test in Prepare request.
Also check one thing that worked before the change.
Save a session before leaving to keep loaded source, your brief, task, and checkpoints. Sessions do not retain folder write permission or a pending AI response. After reopening a session, use Refresh source to choose the same project folder and keep your notes while loading the latest code.
Checkpoint history
No checkpoints yet. Record your test above, even if the change did not work.
Ready for another change? Refresh source to load your latest code, then prepare a new request.