mcp-readwise-bridge
A FastAPI MCP server that exposes your Readwise Reader library to MCP-compatible clients (Claude, VS Code, etc).
GitHubスター
2
ユーザー評価
未評価
お気に入り
0
閲覧数
12
フォーク
0
イシュー
0
MCP Readwise Bridge
A FastAPI MCP server that exposes your Readwise Reader library to MCP-compatible clients (Claude, VS Code, Cursor, Raycast, and more).
Features
- List, retrieve, and update your Readwise Reader documents via a simple HTTP API
- Supports pagination, filtering, and HTML content retrieval
- Handles Readwise rate limits and authentication
- In-memory caching for efficiency
- Compatible with any client that supports the MCP protocol
Quickstart
1. Clone and Install
git clone https://github.com/leonardsellem/mcp-readwise-bridge.git
cd mcp-readwise-bridge
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
2. Configure
Copy .env.example
to .env
and set your Readwise token:
cp .env.example .env
# Edit .env and set READWISE_TOKEN=rw_live_xxx (get your token from https://readwise.io/access_token)
3. Run the Server
uvicorn mcp_reader_bridge.app:api --reload --port 5678
- The server will be available at: http://localhost:5678
- Swagger UI for testing: http://localhost:5678/docs
🚀 One-Click Install for IDEs (VS Code, Cursor, Raycast, Claude, etc.)
Paste the following block into your IDE’s MCP extension or “Add Server” dialog for instant setup:
{
"mcp": {
"inputs": [
{
"type": "promptString",
"id": "readwise_token",
"description": "Readwise Access Token (get from https://readwise.io/access_token)",
"password": true
}
],
"servers": {
"readwise": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"READWISE_TOKEN",
"-p",
"5678:5678",
"ghcr.io/leonardsellem/mcp-readwise-bridge:latest"
],
"env": {
"READWISE_TOKEN": "${input:readwise_token}"
}
}
}
}
}
- This will prompt you for your Readwise token and launch the server in Docker.
- The server will be available at
http://localhost:5678
for your IDE or AI client.
Connecting to Clients
VS Code (with Claude, Cursor, or other MCP plugins)
- Install the MCP extension (e.g., Cursor MCP, Claude VS Code, or your preferred MCP-compatible extension).
- Add a new MCP server in the extension settings:
- Server URL:
http://localhost:5678
- (Optional) Name:
Readwise MCP
- Server URL:
- Authenticate if prompted (the server uses your
.env
token, no extra login needed). - Use the extension’s UI to browse, search, and update your Readwise documents directly from VS Code.
Claude (Anthropic)
- In Claude’s “Connect Knowledge” or “Add Source” dialog, select “Custom MCP server” and enter
http://localhost:5678
. - Claude will now be able to list, retrieve, and update your Readwise documents.
Cursor
- Open Cursor’s command palette and search for “Connect MCP Server”.
- Enter
http://localhost:5678
as the endpoint. - You can now use Cursor’s AI features with your Readwise library.
Raycast
- Install the Raycast MCP extension.
- Add a new MCP server with the URL
http://localhost:5678
. - Raycast will now let you search and interact with your Readwise documents.
Troubleshooting
- 401 Unauthorized: Make sure your
READWISE_TOKEN
is correct and not expired. - 429 Rate Limited: The Readwise API has request limits. Wait a minute and try again.
- Cannot connect: Ensure the server is running and accessible at the port you specified.
- Firewall/Network: If connecting from another device, ensure your firewall allows incoming connections on the server port.
Endpoints
Endpoint | Method | Description |
---|---|---|
/list_documents | GET | List documents (filter, paginate, withContent) |
/get_document/{id} | GET | Retrieve a single document |
/update_document/{id} | PATCH | Update document fields |
FAQ
Q: Is my Readwise data safe?
A: Yes. Your token is stored locally in .env
and never shared with clients. The server only proxies requests.
Q: Can I run this on a remote server?
A: Yes, but you may need to set up HTTPS and firewall rules for security.
Q: Can I use this with other MCP clients?
A: Yes! Any client that supports the MCP protocol can connect.
License
MIT
Links
19
フォロワー
5
リポジトリ
0
Gist
0
貢献数
The Okta MCP Server is a groundbreaking tool built by the team at Fctr that enables AI models to interact directly with your Okta environment using the Model Context Protocol (MCP). Built specifically for IAM engineers, security teams, and Okta administrators, it implements the MCP specification to help work with Okta enitities