GroupDocs.Parser.Mcp

# GroupDocs.Parser MCP Server

MCP server for GroupDocs.Parser — extract text, images, metadata, tables, and barcodes from documents via AI agents. Wraps GroupDocs.Parser for .NET and exposes it as tool calls for Claude, Cursor, VS Code / GitHub Copilot, and other MCP-compatible clients. Supports PDF, DOCX, XLSX, PPTX, HTML, EPUB, MSG, EML, JPG, PNG, TIFF, and 50+ more document and image formats. Exposes ExtractText, ExtractImages, ExtractMetadata, ExtractTables, ExtractBarcodes, and GetDocumentInfo tools.

## Install

Docker-first product - the packed tool exceeds NuGet.org's 250 MB limit, so there is no dnx / NuGet path. Multi-arch image (amd64 + arm64) on GHCR only; pin by replacing :latest with a version tag.

- As a Docker container: `docker run --rm -i -v $(pwd)/documents:/data ghcr.io/groupdocs-parser/parser-net-mcp:latest`

## MCP tools

- ExtractText — extract plain text from a document (whole or per page). Truncates large outputs.
- ExtractImages — save embedded images to storage as `<basename>_image<N>.<ext>` files; returns a saved-path list.
- ExtractMetadata — author / title / dates / EXIF / XMP / IPTC / custom properties as JSON.
- ExtractTables — tables as Markdown (default — renders in chat) or structured JSON (`format='json'`).
- ExtractBarcodes — decoded values + types + page + confidence + angle for each barcode / QR code found.
- GetDocumentInfo — file type + page count + size as JSON, without modifying the file.

## Supported formats

PDF, DOCX, XLSX, PPTX, ODT, RTF, HTML, EPUB, MSG, EML, JPG, PNG, TIFF, BMP, GIF, MP3, MP4, and 50+ other document, image, and media formats.

## Environment variables

- GROUPDOCS_MCP_STORAGE_PATH — base folder for input + output (defaults to current working directory)
- GROUPDOCS_MCP_OUTPUT_PATH — optional, routes output files (used by ExtractImages) to a separate folder
- GROUPDOCS_LICENSE_PATH — path to GroupDocs.Total.lic; omit for evaluation mode (text outputs may be watermarked and other outputs size-limited)

## Licensing

The MCP server is MIT; the underlying GroupDocs engines require a license for production use. Evaluation mode (no license): Text output may include an evaluation watermark, and other outputs may be size-limited. Mount GroupDocs.Total.lic into the container and set GROUPDOCS_LICENSE_PATH. Free 30-day temporary license: https://purchase.groupdocs.com/temporary-license/ - purchase: https://purchase.groupdocs.com/pricing/parser/net

## Key links

- GHCR package: https://github.com/orgs/groupdocs-parser/packages/container/package/parser-net-mcp
- Repository: https://github.com/groupdocs-parser/GroupDocs.Parser.Mcp
- Docker: ghcr.io/groupdocs-parser/parser-net-mcp and docker.io/groupdocs/parser-net-mcp
- License: MIT

## Example prompts for AI agents

- "How many pages does invoice.pdf have, and what format is it?"
- "Extract the text from page 2 of contract.docx."
- "What's the author and creation date of report.xlsx?"
- "Pull the line items table out of invoice.pdf as Markdown."
- "Are there any QR codes in shipping-label.png? If so, what do they decode to?"