Claude2OpenAI APIs to OpenAI API endpoints on the fly router

The Claude2OpenAI is a python server that has the following features:

1. Ability to intercepts the calls from Claude cli in Claude API format.
2. Convert those calls to OpenAI format.
3. Route the converted calls to an OpenAI compatible server.
4. Get the response from the OpenAI compatible server.
5. Convert the Open AI response to Claude API format.
6. Send the response back to the Claude cli.

In the given config.toml file add only the following 2 values and launch the python server. Start using claude like you normally would and voila, everything just works!

• Open AI Base URL
• Open AI Key

Add any other necessary parameters to the configuration file.

Usage

See docs/USAGE.md for setup, configuration, and request/response details.

Claude Code Settings needed for Claude to work with Claude2OpenAI server

Modify your ~/.claude/settings.json to the following.

{
  "env": {
    "ANTHROPIC_API_KEY": "non-empty-string",
    "ANTHROPIC_BASE_URL": "http://127.0.0.1:8080/v1",
    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1",
    "CLAUDE_CODE_ENABLE_TELEMETRY": "0"
  },
  "autoUpdatesChannel": "latest",
  "api_base_url": "http://127.0.0.1:8080/v1",
  "api_key": "any-non-empty-string",
  "apiBaseUrl": "http://127.0.0.1:8080/v1",
  "apiKey": "any-non-empty-string"
}

Export following environment variables from your .basrc/.zshrc shell as needed.

export DISABLE_TELEMETRY="true"
export DISABLE_COST_WARNINGS="true"
export API_TIMEOUT_MS="600000"
unset CLAUDE_CODE_USE_BEDROCK

Raise an issue in github as needed. Feel free to contribute. Source code released under (Apache license 2.0)[https://choosealicense.com/licenses/apache-2.0/].

(c) 2026 Naresh Mehta