JSON Query MCP
@mgraczyk
An MCP server that gives AI agents tools to query large JSON files
概览
What is JSON Query MCP?
An MCP server that enables LLMs implementing the Model Context Protocol to query large JSON files using JSONPath expressions and string-based key/value searching. It is designed for handling JSON data too large to fit in a single LLM context window.
How to use JSON Query MCP?
Run npx json-query-mcp from the command line. For Cursor, add an entry to ~/.cursor/mcp.json specifying "command": "npx" and "args":["<path-to-repo>"].
Key features of JSON Query MCP
- Query JSON files using JSONPath expressions
- Search for keys that match a query string
- Search for values that match a query string
- Handle JSON files larger than 1 million characters
Use cases of JSON Query MCP
- Extract a small subset of data from a massive JSON Swagger definition to generate TypeScript interfaces
- Find specific keys or values in complex or deeply nested JSON documents
- Enable LLMs to work with JSON files that exceed typical context window limits
FAQ from JSON Query MCP
What does JSON Query MCP do?
It provides tools for LLMs to query and search large JSON files without loading the entire file into the model’s context.
How do I install JSON Query MCP?
Run npx json-query-mcp, or configure it as an MCP server in Cursor by adding the command and arguments to mcp.json.
Does JSON Query MCP have any runtime dependencies?
The README does not specify any dependencies beyond Node.js (implied by npx and npm commands).
Can JSON Query MCP handle files larger than 1 million characters?
Yes, the example demonstrates it reading a Swagger definition exceeding 1 million characters and extracting a small portion.
Where does the JSON data live?
The README does not specify; data is presumably passed as a file path or content via the MCP tool interface.
