Pinboard

This MCP Server provides a minimal suite of tools to interact with Pinboard’s API, enabling you to manage bookmarks, tags, and tag suggestions directly from your MCP client. It’s designed for quick, practical queries like reviewing recent bookmarks, analyzing tagging patterns, and refining your tag taxonomy.

How to use

You connect to the Pinboard MCP Server through your MCP client (for example Claude Desktop) and issue natural language requests. You can retrieve bookmarks within a date range, add or update bookmarks, list and rename tags, and fetch tag suggestions. Practical uses include reviewing all bookmarks from a specific period, identifying main topics you saved recently, and analyzing your bookmarking patterns over weeks or months.

How to install

docker pull ghcr.io/vicgarcia/pinboard-mcp:latest

Obtain your Pinboard API token from your Pinboard account settings. The token format is username:1234567890ABCDEF1234567890ABCDEF.

{
  "mcpServers": {
    "pinboard": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "PINBOARD_TOKEN=your-username:your-api-token",
        "ghcr.io/vicgarcia/pinboard-mcp:latest"
      ]
    }
  }
}

To run a locally built image instead of the latest published one, build and run a local tag as shown below, then configure CLAUDE to use the local image. You must replace the token with your actual value.

docker build -t pinboard-mcp:local .

To use the local build in CLAUDE Desktop:

{
  "mcpServers": {
    "pinboard": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "PINBOARD_TOKEN=your-username:your-api-token",
        "pinboard-mcp:local"
      ]
    }
  }
}

Additional notes

Security: keep your Pinboard API token confidential. Do not expose tokens in shared configurations. If you use the local build flow, ensure the token is supplied securely in your environment.

Tools and capabilities

This MCP Server exposes a focused set of tools to work with Pinboard data. You can retrieve and filter bookmarks, add new bookmarks, update existing ones, inspect and rename tags, and obtain tag suggestions for URLs.

Available tools

get_bookmarks

Retrieve bookmarks within a date range with optional tag filtering and a limit. Supports start_date, end_date, tags, and limit.

add_bookmark

Create a new bookmark with url, title, optional description, tags, privacy, and to-read flag.

update_bookmark

Update an existing bookmark by URL, including title, description, tags, privacy, and to-read status.

get_tags

Retrieve all tags with usage counts, sorted by frequency.

rename_tag

Rename an existing tag across all bookmarks, normalizing to lowercase and ensuring non-empty values.

suggest_tags

Get tag suggestions for a URL, including popular and personalized recommended tags.