mcp-oauth2-token-provider
Acts as an authenticated proxy between AI IDEs and remote MCP endpoints, handling OAuth2 client_credentials token generation and transparently injecting bearer tokens into all proxied requests.
README
mcp-oauth2-token-provider
A generic MCP (Model Context Protocol) server that acts as an authenticated proxy between AI IDEs and remote MCP endpoints. Handles OAuth2 client_credentials token generation internally and transparently injects bearer tokens into all proxied requests.
Features
- Transparent OAuth2 proxy — acquires, caches, and auto-refreshes tokens
- Dynamic tool discovery — no hardcoded tool schemas; discovers tools from the remote MCP endpoint
- Auto-retry on 401 — if a token expires mid-session, refreshes and retries once
- Startup health check — verifies remote MCP is reachable before connecting; fails fast with clear error messages
- Token persistence — caches tokens to disk at
~/.oauth2-token-provider/token.jsonto survive process restarts - Zero configuration in code — all settings via environment variables
- Supports SSE and JSON responses — handles both Streamable HTTP and plain JSON-RPC from the remote
Installation
npx mcp-oauth2-token-provider
Or install globally:
npm install -g mcp-oauth2-token-provider
Usage with Kiro / VS Code MCP
Add to your .kiro/settings/mcp.json (or equivalent):
{
"mcpServers": {
"my-remote-server": {
"command": "npx",
"args": ["mcp-oauth2-token-provider"],
"env": {
"REMOTE_MCP_URL": "https://your-server.com/mcp",
"OAUTH2_TOKEN_URL": "https://your-sso.com/token.oauth2",
"OAUTH2_CLIENT_ID": "your-client-id",
"OAUTH2_CLIENT_SECRET": "your-client-secret"
}
}
}
}
Environment Variables
| Variable | Required | Description |
|---|---|---|
REMOTE_MCP_URL |
Yes | The remote MCP endpoint URL to proxy to |
OAUTH2_TOKEN_URL |
Yes | OAuth2 token endpoint (client_credentials grant) |
OAUTH2_CLIENT_ID |
Yes | OAuth2 client ID |
OAUTH2_CLIENT_SECRET |
Yes | OAuth2 client secret |
OAUTH2_BASIC_USERNAME |
No | Basic auth username (for dual-credential SSO) |
OAUTH2_BASIC_PASSWORD |
No | Basic auth password (for dual-credential SSO) |
How It Works
AI IDE ←stdio→ [mcp-oauth2-token-provider] ←HTTP+Bearer→ [Remote MCP Server]
│
├─ Acquires OAuth2 token (client_credentials)
├─ Caches in memory + ~/.oauth2-token-provider/token.json
├─ Injects Bearer header on every request
├─ Auto-refreshes on expiry (60s safety margin)
└─ Retries once on 401 with fresh token
- On startup, verifies the remote MCP endpoint is reachable (sends MCP
initialize) - If remote is unreachable, exits with error (IDE shows "Failed" status)
- If healthy, connects to IDE via stdio and discovers tools from the remote
- All tool calls are proxied transparently with the bearer token
Token Caching
Tokens are cached in two layers:
- In-memory — fastest, lost on process restart
- On disk —
~/.oauth2-token-provider/token.json, survives restarts
The token is refreshed when:
- It's within 60 seconds of expiry
- A 401 response is received from the remote
Dual-Credential SSO (Optional)
Some SSO providers require both:
- A
client_id+client_secretin the request body - A Basic Auth header with service account credentials
Set OAUTH2_BASIC_USERNAME and OAUTH2_BASIC_PASSWORD to enable this pattern.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.