Jina AI

Lightweight MCP server to fetch web pages as Markdown via r.jina.ai, with hashing, tokenization, and caching for efficient LLM access.

other

1

GitHub Stars

other

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "jswords91-jina-ai-mcp": {
      "command": "uv",
      "args": [
        "run",
        "src/jina-mcp/server.py"
      ]
    }
  }
}

This lightweight MCP server fetches and serves Markdown versions of web pages via r.jina.ai, enabling language models to access clean, token-aware web content with minimal overhead. It exposes a stable MCP endpoint that you can integrate with LLM clients to retrieve web content as Markdown for downstream processing.

How to use

You use this MCP server by connecting an MCP client to the local or remote MCP endpoint. The server exposes a simple, token-friendly way to fetch any URL as Markdown, with built-in auto-hashing, tokenization, and caching. To test locally, run the start command below and then point your MCP client at the server to fetch Markdown-formatted pages as needed.

How to install

Prerequisites you need before starting: Node.js and npm must be installed on your machine. You will run a runtime command that executes the MCP server script.

Ensure Node.js and npm are installed on your system. You can verify with:

ode -v
npm -v

Start the MCP server locally using the provided runtime command. This runs the server script that serves Markdown versions of web pages via the MCP interface.

npx @modelcontextprotocol/inspector uv run src/jina-mcp/server.py

If you want to include the MCP server in a configuration file, you can use the following snippet. It defines an MCP server named jina_mcp that runs with the uv runtime and executes the server script.

{
  "mcpServers": {
    "jina_mcp": {
      "command": "uv",
      "args": [
        "run",
        "src/jina-mcp/server.py"
      ]
    }
  }
}

Additional notes and configuration

This MCP server is designed to fetch any URL as Markdown, auto-hash content, tokenize for efficient token usage, and cache results for faster repeated access. It assigns stable URIs for resource lookup, enabling reliable integration with various MCP clients.

Available tools

fetchMarkdown

Fetch any URL and convert the content to Markdown for downstream processing by LLMs.

hashCache

Auto-hashes fetched content to generate stable cache entries and help with deduplication.

tokenizeContent

Tokenizes fetched Markdown content to optimize token usage when sending data to LLMs.