Fetch JSONPath MCP

How to use

Install and run the MCP server, then use its tools to fetch and process data from URLs. You can extract specific JSON data with JSONPath patterns, fetch text content from web pages with intelligent extraction, and run batch requests to handle multiple URLs at once. The server automatically uses realistic browser headers to avoid blocks and to improve access to targets.

How to install

Prerequisites: ensure you have the MCP runtime tooling available in your environment. The server is designed to be driven by an MCP client and supports local execution via a command-line tool.

# 1) Synchronize dependencies
uv sync

# 2) (Optional) Start a demo server to test against
# Install demo server dependencies
uv add fastapi uvicorn

# Start demo server on port 8080
uv run demo-server

# 3) Run the MCP server locally
uv run fetch-jsonpath-mcp

If you prefer to connect this MCP server via a config, you can place the following snippet in your MCP client configuration to run the server locally with the UVX runtime.

{
  "mcpServers": {
    "fetch-jsonpath-mcp": {
      "command": "uvx",
      "args": ["fetch-jsonpath-mcp"]
    }
  }
}

Configuration and runtime options

You can customize how the MCP server fetches content by tweaking environment variables. The following variables are supported:

export JSONRPC_MCP_TIMEOUT=30
export JSONRPC_MCP_VERIFY=false
export JSONRPC_MCP_FOLLOW_REDIRECTS=true
export JSONRPC_MCP_HEADERS='{"Authorization": "Bearer token"}'
export JSONRPC_MCP_PROXY="http://proxy.example.com:8080"

Defaults are applied when a variable is not provided. The server uses realistic browser headers by default to avoid access blocks and to improve content extraction. Custom headers replace defaults when there are conflicts.

Available tools overview

You can perform data extraction and content retrieval with these tools. Each tool is designed to work with JSON data or web pages and supports common HTTP methods.

Troubleshooting and notes

If you encounter access blocks, verify that the browser-like headers are being sent and that redirects are followed if needed. Ensure your environment variables are set correctly and that the server process is running with the expected runtime (UVX) and command arguments.

Examples you can try

# Example: fetch JSON with a pattern
# This will extract data matching the JSONPath pattern data[*].name
# from a target URL
{
  "name": "fetch-json",
  "arguments": {
    "url": "http://localhost:8080",
    "pattern": "data[*].name",
    "method": "GET"
  }
}

Demo server data and endpoints

A local demo server can provide sample JSON data to test extraction and processing.

Notes on multi-step flows

If you perform installation, build, and run steps, ensure you reach the final start command to keep the server running. The example start command shown above runs the MCP server locally via UVX.

fetch-json

Extract JSON data using JSONPath patterns with support for all HTTP methods.

fetch-text

Fetch web content with intelligent text extraction and default Markdown formatting.

batch-fetch-json

Process multiple URLs with different JSONPath patterns concurrently.

batch-fetch-text

Fetch content from multiple URLs with intelligent text extraction.