Tool Integrations
Guides for integrating MyTokenGate with popular AI tools and platforms.
Model Context Protocol
MCP
How MCP-capable clients use MyTokenGate as the model backend while running MCP servers on your side.
AI Code Editors
Claude Code
AI-powered terminal coding assistant from Anthropic.
Cursor
AI-first code editor with built-in AI assistance.
Cline
Autonomous coding agent for VS Code.
Codex CLI
OpenAI’s official terminal coding assistant.
Aider
AI pair programming in the terminal; reads and writes your code and commits to Git.
Hermes
Self-improving AI agent framework by Nous Research.
OpenClaw
Open-source multi-platform personal AI assistant framework.
Chat Applications
Chatbox
Cross-platform desktop chat application.
ChatHub
Browser extension for multiple AI chatbots.
Cherry Studio
Cross-platform desktop AI client with custom OpenAI-compatible providers.
SillyTavern
AI chat frontend with character customization.
Platforms
Dify
LLM application development platform.
Base URL Configuration Guide
Different tools require different Base URL formats. Refer to the table below:
| Tool | Base URL | Notes |
|---|---|---|
| OpenAI SDK | https://gateway.mytokengate.com/v1 | SDK appends /chat/completions |
| Anthropic SDK | https://gateway.mytokengate.com/v1 | SDK appends /messages |
| Claude Code | https://gateway.mytokengate.com | Tool auto-adds /v1/messages |
| Continue (VSCode) | https://gateway.mytokengate.com | Tool auto-adds version path |
| Cline (VSCode) | https://gateway.mytokengate.com | Tool auto-adds version path |
| LangChain | https://gateway.mytokengate.com/v1 | Passes directly to SDK |
| Cursor IDE | https://gateway.mytokengate.com/v1 | OpenAI-compatible endpoint |
| Codex CLI | https://gateway.mytokengate.com/v1 | Custom provider in config.toml |
| Aider | https://gateway.mytokengate.com/v1 | Set OPENAI_API_BASE, prefix model with openai/ |
| Cherry Studio | https://gateway.mytokengate.com/v1 | OpenAI-type custom provider |
| Hermes Agent | https://gateway.mytokengate.com/v1 | Use custom provider config |
| OpenClaw | https://gateway.mytokengate.com/v1 | Custom provider config |
| curl / HTTP | https://gateway.mytokengate.com/v1 | Use full path in request |
Configuration Rules
- If the tool mentions “OpenAI-compatible” or uses official SDKs → include
/v1 - If the tool is an AI assistant plugin (Continue, Cline, Claude Code) → omit
/v1
Protocol Selection Guide
MyTokenGate Gateway supports multi-protocol access with internal protocol conversion. Using the model’s native protocol provides the best compatibility:
| Model Series | Native Protocol | Recommended Configuration |
|---|---|---|
| Claude | Anthropic | Use Anthropic SDK or Anthropic provider settings |
| GPT | OpenAI | Use OpenAI SDK or OpenAI Compatible settings |
| Gemini | OpenAI | Use OpenAI Compatible settings |
| DeepSeek | OpenAI | Use OpenAI Compatible settings |
| Chinese Models | OpenAI | Use OpenAI Compatible settings |
Why choose native protocol?
- Better feature compatibility (e.g., Claude’s extended thinking)
- Less parameter conversion overhead
- More complete error messages
- More stable performance
Configuration Examples
Each integration has its own page with complete setup steps — follow the links above. The examples here cover the generic SDKs plus Continue, which does not have a dedicated page yet.
Continue (VSCode)
Add to your config.json:
{
"models": [{
"title": "MyTokenGate Claude",
"provider": "anthropic",
"apiBase": "https://gateway.mytokengate.com",
"apiKey": "tg-your-api-key"
}]
}Note: Use apiBase without /v1
OpenAI SDK
from openai import OpenAI
client = OpenAI(
base_url="https://gateway.mytokengate.com/v1",
api_key="tg-your-api-key"
)Anthropic SDK
from anthropic import Anthropic
client = Anthropic(
base_url="https://gateway.mytokengate.com/v1",
api_key="tg-your-api-key"
)