{
  "markdown": "# mcp-pymupdf\n\n<!-- mcp-name: io.github.penny4nonsense/mcp-pymupdf -->\n\nAn MCP server for reading PDFs using PyMuPDF. Supports local files and URLs.\n\n## Installation\n\n```bash\npip install mcp-pymupdf\n```\n\n## Tools\n\n### `read_pdf`\nExtract text from a PDF file or URL.\n\n```json\n{\n  \"source\": \"/path/to/file.pdf\",\n  \"max_pages\": 50,\n  \"max_bytes\": 200000\n}\n```\n\n### `get_page`\nExtract text from a specific page range.\n\n```json\n{\n  \"source\": \"https://example.com/paper.pdf\",\n  \"start_page\": 3,\n  \"end_page\": 8\n}\n```\n\n### `get_metadata`\nRetrieve PDF metadata without reading full text.\n\n```json\n{\n  \"source\": \"/path/to/file.pdf\"\n}\n```\n\n## Usage with Claude Desktop\n\nAdd to your Claude Desktop config:\n\n```json\n{\n  \"mcpServers\": {\n    \"pdf\": {\n      \"command\": \"mcp-pymupdf\"\n    }\n  }\n}\n```\n\n## Notes\n\n- Scanned or image-based PDFs cannot be read without OCR\n- Mathematical notation may render as plain text approximations\n- URLs are downloaded to a temp file and deleted after reading\n\n## License\n\nMIT",
  "bytes": 995,
  "sha": "d2a00a97d05fce8bfba00151580add70b98ab621890d5b65ea7937d17f637f54",
  "repo_slug": "penny4nonsense/mcp-pymupdf",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_penny4nonsense_mcp_pymupdf_c9ddabec/readme"
}