prescription
The OpenAI-compatible interface via the Gemini-CLI-2-API allows for seamless integration with existing toolchains. The steps are described below:
- Deployment of local services: First by
npm installInstall the dependencies and use thenode openai-api-server.jsStart the service (default port 8000) - Configuring the Terminal Address: Change the API endpoint address of existing tools to
http://localhost:8000/v1 - Maintaining the request format: Continue to use the OpenAI standard JSON request body (e.g.
/v1/chat/completions(Endpoints) - Authentication Adaptation: If the original tool uses an API key, start the service via the
--api-keyParameterize the same key
Advantages are: 1) Automatic conversion of request/response format 2) Support streaming 3) Retain the original function call method. Applicable to LangChain, AutoGPT and other frameworks.
This answer comes from the articleGemini-CLI-2-API: Converting the Gemini CLI to an OpenAI-compatible Native API ServiceThe
May not be reproduced without permission:AI productivity tools " How to integrate Gemini models in an existing OpenAI toolchain and avoid code changes?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil