Codex Bridge
Connecting providers
Workspace / Request monitor

Every request, in view.

Codex and Claude. Follow each prompt to its response, live.

Total requests0Saved in this workspace
In progress0Connecting
Average first text—Includes time in queue
Completed0No requests yet

Start with a prompt.

Your response will stream here.
Follow each stage in the activity panel.

Try a quick request

New request

Choose Codex or Claude and send a prompt.

Lower reasoning usually responds sooner. Each request uses a fresh conversation in your selected CLI account.

Ctrl + Enter to send

Call it from anywhere in your app.

One HTTP endpoint. A complete response or a live stream.

Get a complete response

Stream the response
Add "stream": true to the JSON body. The endpoint returns server-sent events: response.delta carries text chunks; request.finished contains the final request and output_text.
Track an asynchronous request
POST /api/requests returns a request ID immediately. Read it at GET /api/requests/{id} or subscribe to GET /api/requests/{id}/events. Cancel with POST /api/requests/{id}/cancel and an empty JSON body.
This is a Codex Bridge API, not a drop-in replacement for the OpenAI API. Requests continue if a client disconnects. Use the cancellation endpoint to stop them.
Download API schema
  
Connect to your bridge
This server requires a bridge API token.
Stored in this tab only