AIClient-2-API/README-EN.md

Gemini-CLI-2-API 🚀

A powerful proxy that unifies multiple large model APIs (Gemini, OpenAI, Claude...) into a local OpenAI-compatible interface.

中文 | English

GeminiCli2API is a versatile and lightweight API proxy designed for maximum flexibility and ease of use. It uses a Node.js HTTP server to transform various backend APIs, such as Google Gemini (CLI authorized), OpenAI, Claude, and Kiro, into a standard OpenAI format interface. The project adopts modern modular architecture with strategy and adapter patterns, complete test coverage and health check mechanisms, ready to use out-of-the-box—simply run npm install and it's good to go. By easily switching the model provider in the configuration file, you can enable any OpenAI-compatible client or application to seamlessly use different large model capabilities through a single API address, completely eliminating the hassle of maintaining multiple configurations and dealing with incompatible interfaces.

---

💡 Core Advantages

• ✅ Unified Access to Multiple Models: One interface for Gemini, OpenAI, Claude, Kimi K2, GLM-4.5, and other latest models. Freely switch between different model service providers with simple startup parameters or request headers.
• ✅ Break Through Official Limits: By supporting authorization via the Gemini CLI's OAuth method, it effectively bypasses the rate and quota limits of the official free API, allowing you to enjoy higher request quotas and usage frequency.
• ✅ Break Through Client Limits: Kiro API mode supports free use of Claude Sonnet 4 model.
• ✅ Seamless OpenAI Compatibility: Provides an interface fully compatible with the OpenAI API, allowing your existing toolchains and clients (like LobeChat, NextChat, etc.) to access all supported models at zero cost.
• ✅ Enhanced Controllability: With powerful logging features, you can capture and record all request prompts, which is convenient for auditing, debugging, and building private datasets.
• ✅ Extremely Easy to Extend: Thanks to the new modular and strategy pattern design, adding a new model service provider has never been easier.
• ✅ Complete Test Coverage: Provides comprehensive integration and unit tests to ensure the stability and reliability of all API endpoints and functions.

---

📝 Project Architecture

Leaving behind the simple structure of the past, we have introduced a more professional and extensible design pattern to completely transform the project:

• src/api-server.js: 🚀 Project Startup Entry

  • As the project's commander-in-chief, it is responsible for starting and managing the entire HTTP service, parsing command-line arguments, and loading all configurations.

• src/adapter.js: 🔌 Service Adapter

  • Adopts the classic adapter pattern to create a unified interface for each AI service (Gemini, OpenAI, Claude). No matter how the backend service changes, the calling method remains consistent for the main service.

• src/provider-strategies.js: 🎯 Provider Strategy Factory

  • Implements the strategy factory pattern, providing unified strategy interfaces for each API protocol (such as OpenAI, Gemini, Claude). These strategies accurately handle request parsing, response formatting, model name extraction and other details under the protocol, ensuring perfect conversion between protocols.

• src/provider-strategy.js: 🎯 Strategy Base Class

  • Defines the basic interface and common methods for all provider strategies, including core functions such as system prompt management and content extraction.

• src/convert.js: 🔄 Format Conversion Center

  • This is the core of the magic that makes "everything OpenAI-compatible." It is responsible for accurate and lossless data conversion between different API protocol formats.

• src/common.js: 🛠️ Common Utility Library

  • Stores shared constants, utility functions, and common handlers for the project, making the code cleaner and more efficient.

• src/gemini/, src/openai/, src/claude/: 📦 Provider Implementation Directories

  • Each directory contains the core logic, API calls, and strategy implementations for the corresponding service provider, with a clear structure that makes it easy for you to add more new service providers in the future. Among them, src/openai/openai-kiro.js provides a special implementation for the Kiro API.

• tests/: 🧪 Test Directory

  • Contains a complete integration test suite covering all API endpoints, authentication methods, and error handling scenarios to ensure project stability and reliability.

---

🔧 Usage Instructions

• MCP Support: While the built-in command functions of the original Gemini CLI are not available, this project perfectly supports MCP (Model Context Protocol) and can work with MCP-compatible clients for more powerful functionality extensions.
• Multimodal Capabilities: Supports multimodal inputs such as images and documents, providing you with a richer interactive experience.
• Latest Model Support: Supports the latest Kimi K2 and GLM-4.5 models. Simply configure the corresponding OpenAI or Claude compatible interfaces in config.json to use them.
• Kiro API: Using Kiro API requires Download Kiro client and completing authorized login to generate kiro-auth-token.json. Recommended for use with Claude Code for the best experience. .

---

🛠️ Key Features

General Features

• 🔐 Smart Authentication & Token Renewal: For services that require OAuth (like gemini-cli-oauth), the first run will guide you through browser authorization and can automatically refresh the token.
• 🛡️ Multiple Authentication Methods: Supports Authorization: Bearer <key>, x-goog-api-key, x-api-key request headers, and URL query parameters for authentication.
• ⚙️ Highly Configurable: Flexibly configure the listening address, port, API key, model provider, and log mode via the config.json file or command-line arguments.
• 📜 Fully Controllable Logging System: Can output timestamped prompt logs to the console or a file, and display the remaining token validity period.
• 🏥 Health Check Mechanism: Provides /health endpoint for service status monitoring, returning service health status and current configuration information.

OpenAI Compatible Interface (/v1/...)

• 🌍 Perfect Compatibility: Implements the core /v1/models and /v1/chat/completions endpoints.
• 🔄 Automatic Format Conversion: Internally and seamlessly converts requests/responses between different model formats and the OpenAI format.
• 💨 Streaming Support: Fully supports OpenAI's streaming responses ("stream": true), providing a typewriter-like real-time experience.

---

📦 Installation Guide

1. Prerequisites:

   • Please ensure you have Node.js installed (recommended version >= 20.0.0).
   • This project already includes package.json and sets {"type": "module"}, so you don't need to create it manually.

2. Install Dependencies: After cloning this repository, execute the following in the project root directory:

   npm install
   

   This will automatically install all necessary dependencies.

---

🚀 Quick Start

1. Configuration File (config.json)

We recommend using the config.json file to manage your configurations, which is clearer than lengthy command-line arguments.

First, manually create a config.json file and fill in your configuration information.

{
    "REQUIRED_API_KEY": "123456",
    "SERVER_PORT": 3000,
    "HOST": "localhost",
    "MODEL_PROVIDER": "gemini-cli-oauth",
    "OPENAI_API_KEY": "sk-your-openai-key",
    "OPENAI_BASE_URL": "https://api.openai.com/v1",
    "CLAUDE_API_KEY": "sk-ant-your-claude-key",
    "CLAUDE_BASE_URL": "https://api.anthropic.com/v1",
    "PROJECT_ID": "your-gcp-project-id",
    "PROMPT_LOG_MODE": "console"
}

2. Configuration Parameter Details

The following are all the supported parameters in the config.json file and their detailed descriptions:

Parameter Name	Type	Description	Default/Optional Values
REQUIRED_API_KEY	string	The key used to protect your API service. Clients must provide this key when making requests.	Any string, defaults to "123456"
SERVER_PORT	number	The port number the server listens on.	Any valid port number, defaults to 3000
HOST	string	The host address the server listens on. localhost only allows local access, 0.0.0.0 allows LAN or public network access.	Defaults to "localhost"
MODEL_PROVIDER	string	Specifies the backend model service provider to use. This is a core configuration that determines which platform API requests will be forwarded to.	Optional values: "gemini-cli-oauth", "openai-custom", "claude-custom", "openai-kiro-oauth"
OPENAI_API_KEY	string	When MODEL_PROVIDER is openai-custom, you need to provide your OpenAI API key.	null
OPENAI_BASE_URL	string	When MODEL_PROVIDER is openai-custom, you can specify an OpenAI-compatible API address.	Defaults to "https://api.openai.com/v1"
CLAUDE_API_KEY	string	When MODEL_PROVIDER is claude-custom, you need to provide your Claude API key.	null
CLAUDE_BASE_URL	string	When MODEL_PROVIDER is claude-custom, you can specify a Claude-compatible API address.	Defaults to "https://api.anthropic.com/v1"
KIRO_OAUTH_CREDS_BASE64	string	(Kiro API mode) The Base64 encoded string of your Kiro OAuth credentials.	null
KIRO_OAUTH_CREDS_FILE_PATH	string	(Kiro API mode) The path to your Kiro OAuth credentials JSON file.	null
GEMINI_OAUTH_CREDS_BASE64	string	(Gemini-CLI mode) The Base64 encoded string of your Google OAuth credentials.	null
GEMINI_OAUTH_CREDS_FILE_PATH	string	(Gemini-CLI mode) The path to your Google OAuth credentials JSON file.	null
PROJECT_ID	string	(Gemini-CLI mode) Your Google Cloud project ID.	null
SYSTEM_PROMPT_FILE_PATH	string	The path to an external file for loading system prompts.	Defaults to "input_system_prompt.txt"
SYSTEM_PROMPT_MODE	string	The application mode for system prompts. overwrite will override the client's prompt, append will append to the end of the client's prompt.	Optional values: "overwrite", "append"
PROMPT_LOG_MODE	string	The logging mode for requests and responses. none does not log, console prints to the console, file saves to a log file.	Optional values: "none", "console", "file"
PROMPT_LOG_BASE_NAME	string	When PROMPT_LOG_MODE is file, the base name for the generated log files.	Defaults to "prompt_log"
REQUEST_MAX_RETRIES	number	The maximum number of times to automatically retry when an API request fails.	Defaults to 3
REQUEST_BASE_DELAY	number	The base delay time (in milliseconds) between automatic retries. The delay will increase after each retry.	Defaults to 1000

3. Start the Service

• Start with config.json (recommended)

  node src/api-server.js
  

• Start with command-line arguments (will override same-name configurations in config.json)
  • Start OpenAI proxy:

    node src/api-server.js --model-provider openai-custom --openai-api-key sk-xxx
    

  • Start Claude proxy:

    node src/api-server.js --model-provider claude-custom --claude-api-key sk-ant-xxx
    

  • Start Kiro API proxy:

    node src/api-server.js --model-provider claude-kiro-oauth
    

  • Listen on all network interfaces and specify port and key (for Docker or LAN access)

    node src/api-server.js --host 0.0.0.0 --port 8000 --api-key your_secret_key
    

For more startup parameters, please refer to the comments at the top of the src/api-server.js file.

---

4. Call the API

Hint: If you are using this in an environment where you cannot directly access Google/OpenAI/Claude/Kiro services, please set up a global HTTP/HTTPS proxy for your terminal first.

All requests use the standard OpenAI format.

• Health Check

  curl http://localhost:3000/health
  

• List Models

  curl http://localhost:3000/v1/models \
    -H "Authorization: Bearer 123456"
  

• Generate Content (Non-streaming)

  curl http://localhost:3000/v1/chat/completions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer 123456" \
    -d '{
      "model": "gemini-2.5-flash",
      "messages": [
        {"role": "system", "content": "You are a cat named Neko."},
        {"role": "user", "content": "Hello, what is your name?"}
      ]
    }'
  

• Stream Generate Content

  curl http://localhost:3000/v1/chat/completions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer 123456" \
    -d '{
      "model": "claude-3-opus-20240229",
      "messages": [
        {"role": "user", "content": "Write a five-line poem about the universe"}
      ],
      "stream": true
    }'
  

---

🌟 Special Usage & Advanced Tips

• 🔌 Connect to Any OpenAI Client: This is the basic feature of this project. Point the API address of any application that supports OpenAI (like LobeChat, NextChat, VS Code extensions, etc.) to this service (http://localhost:3000) to seamlessly use all configured models.

• 🔍 Centralized Request Monitoring & Auditing: Set "PROMPT_LOG_MODE": "file" in config.json to capture all requests and responses and save them to a local log file. This is crucial for analyzing, debugging, and optimizing prompts, and even for building private datasets.

• 💡 Dynamic System Prompts:

  • By setting SYSTEM_PROMPT_FILE_PATH and SYSTEM_PROMPT_MODE in config.json, you can control the behavior of system prompts more flexibly.
  • Supported Modes:
    • override: Completely ignores the client's system prompt and forces the use of the content from the file.
    • append: Appends the content of the file to the end of the client's system prompt to supplement rules.
  • This allows you to set unified base instructions for different clients while allowing individual applications for personalized extensions.

• 🛠️ Foundation for Secondary Development:

  • Add New Models: Simply create a new provider directory under src, implement the ApiServiceAdapter interface and the corresponding strategies, and then register it in adapter.js and common.js.
  • Response Caching: Add caching logic for frequently repeated questions to reduce API calls and improve response speed.
  • Custom Content Filtering: Add keyword filtering or content review logic before requests are sent or returned to meet compliance requirements.

---

📄 License

This project is licensed under the GNU General Public License v3 (GPLv3). For details, please see the LICENSE file in the root directory.

🙏 Acknowledgements

The development of this project was greatly inspired by the official Google Gemini CLI, and referenced some code implementations from Cline 3.18.0's gemini-cli.ts. I would like to express my sincere gratitude to the official Google team and the Cline development team for their excellent work!