Run OpenAI Codex on Cloudflare Sandboxes! This example shows a basic setup that does the following:

• The worker accepts POST requests that include a repository URL and a task description
• The worker spawns a sandbox, clones the repository and runs codex exec in non-interactive mode with the provided task
• Codex will edit all necessary files and return when done
• The Worker will return a response with the output logs from Codex and the diff left on the repo.

The OpenAI credential never enters the container. The worker subclasses Sandbox with interceptHttps = true and registers outboundByHost handlers for api.openai.com and chatgpt.com that swap a placeholder header for the real secret on its way out. Inside the container, codex only sees OPENAI_API_KEY=proxy-injected (or a placeholder ~/.codex/auth.json) -- enough to make it pick the right auth mode, but useless if leaked.

Internet access is disabled inside the container, bar the OpenAI, ChatGPT and GitHub APIs; this lets codex exec run headlessly without needing to ask for permissions.

Copy .dev.vars.example to .dev.vars and fill in one of:

# Pay-per-token
OPENAI_API_KEY=<your-api-key>

# OR ChatGPT subscription -- generate by running `codex login`, then paste
# the contents of ~/.codex/auth.json into the variable as a single-line JSON.
CODEX_AUTH_JSON=<your-auth-json>

If both are set, the API key wins.

For production, set secrets with:

wrangler secret put OPENAI_API_KEY
# or
wrangler secret put CODEX_AUTH_JSON

Note on CODEX_AUTH_JSON: Treat the file like a password -- it contains access tokens. Codex will refresh stale tokens during a run, but in this stateless setup the refreshed bundle isn't persisted back to your secret store. For long-lived deployments, prefer the API key route or follow the CI/CD auth guide to persist the refreshed file between runs.

curl -X POST http://localhost:8787/ \
  -H 'Content-Type: application/json' \
  -d '{"repo": "https://github.com/owner/repo", "task": "fix the typo in README.md"}'

Response:

{
  "logs": "...",
  "diff": "..."
}

Happy hacking!