GroupDocs.Comparison MCP Server
Enables AI agents to compare documents, analyze changes, and retrieve document info using GroupDocs.Comparison. Supports various formats and includes licensing for production use.
README
GroupDocs.Comparison MCP Server
MCP server that exposes GroupDocs.Comparison as AI-callable tools for Claude, Cursor, GitHub Copilot, and other MCP agents.
<!-- TODO(manual): record 30s demo per improvement-plan M1 -->
<!--
-->
Installation
<!-- install-buttons:start - generated by install/generate-install-links.ps1; do not edit by hand -->
One-click installs pre-fill every supported setting: edit the placeholder documents folder after install (output defaults to the same folder); an empty GROUPDOCS_LICENSE_PATH runs in evaluation mode - point it at your license to lift limits.
More clients - ready-made configs for Claude Code, Codex CLI, Visual Studio 2022, Cursor, Windsurf, Cline, and JetBrains Rider live in
install/generated/. <!-- install-buttons:end -->
Requires .NET 10 SDK.
Run directly with dnx (recommended — no install step):
dnx GroupDocs.Comparison.Mcp --yes
Pulls the latest stable release on every invocation. To pin to a specific
version (recommended for shared configs and CI), append @<version>:
dnx GroupDocs.Comparison.Mcp@26.7.4 --yes
Or install as a global dotnet tool:
dotnet tool install -g GroupDocs.Comparison.Mcp
groupdocs-comparison-mcp
Or run via Docker:
docker run --rm -i \
-v $(pwd)/documents:/data \
ghcr.io/groupdocs-comparison/comparison-net-mcp:latest
Native prerequisites
The underlying GroupDocs engine uses System.Drawing (GDI+) for some
operations. When you run the server natively (via dnx or the global
dotnet tool) on Linux or macOS, install the native libgdiplus library first:
| Platform | Setup |
|---|---|
| Windows | Nothing — GDI+ is built into the OS. |
| Linux | sudo apt-get install -y libgdiplus libfontconfig1 ttf-mscorefonts-installer |
| macOS | brew install mono-libgdiplus |
| Docker | Nothing — the image already bundles libgdiplus. |
Skipping this on Linux/macOS surfaces as DllNotFoundException: libgdiplus in
the tool response. The simplest zero-setup option on Linux/macOS is the
Docker image.
Available MCP Tools
| Tool | Description |
|---|---|
Compare |
Compares two documents (source vs target) and produces a marked-up result file, plus a change-count summary and a structured JSON list of the changes (type, page, the changed fragment, surrounding source/target text, table cell, style changes). Use when the user wants the rendered diff document. Supports PDF, Word, Excel, PowerPoint, ODT, RTF, TXT, HTML, and 30+ more formats; optional sourcePassword / targetPassword cover protected documents. |
AnalyzeChanges |
Returns the differences between two documents as structured data only — the same JSON change list as Compare, but without rendering or saving a result file. Cheaper than Compare; use when the user wants to summarize, explain, or reason about what changed rather than obtain the marked-up file. Optional sourcePassword / targetPassword. |
GetDocumentInfo |
Inspects a single source document and returns file type, page count, file size, and per-page dimensions as JSON — without performing a comparison. Useful as a pre-flight check before deciding whether to compare or which formats to expect. Optional password for protected documents. |
Example prompts for AI agents
Once the server is wired up to your MCP client (Claude Desktop, Cursor, VS Code Copilot, …), try:
Compare old.pdf and new.pdf and save the marked-up result.
Diff contract-v1.docx against contract-v2.docx and tell me the change count.
What changed between proposal-v1.docx and proposal-v2.docx? Summarize the edits.
Did the prices change between budget-q1.xlsx and budget-q2.xlsx? Which cells?
How many pages does report.pdf have? Who's the author?
Inspect /docs/legal-brief.pdf — what's the file type and page count?
The client picks Compare when the user wants the rendered diff file,
AnalyzeChanges when they only want to know what changed (summary /
analysis, no file), and GetDocumentInfo for inspection-only questions.
Licensing
The MCP server itself is MIT; the underlying GroupDocs.Comparison engine requires a license for production use. Without one the server runs in evaluation mode:
- Comparison output may carry evaluation watermarks / notices.
- A single server process can open at most 15 documents; further calls fail with an evaluation-cap error until the server restarts.
To lift the limits, point GROUPDOCS_LICENSE_PATH at your GroupDocs.Total.lic:
Configuration
| Variable | Description | Default |
|---|---|---|
GROUPDOCS_MCP_STORAGE_PATH |
Base folder for input and output files | current directory |
GROUPDOCS_MCP_OUTPUT_PATH |
(Optional) separate folder for output files | GROUPDOCS_MCP_STORAGE_PATH |
GROUPDOCS_LICENSE_PATH |
Path to GroupDocs license file | (evaluation mode) |
Usage with Claude Desktop
{
"mcpServers": {
"groupdocs-comparison": {
"type": "stdio",
"command": "dnx",
"args": ["GroupDocs.Comparison.Mcp", "--yes"],
"env": {
"GROUPDOCS_MCP_STORAGE_PATH": "/path/to/documents"
}
}
}
}
To pin to a specific version, replace
"GroupDocs.Comparison.Mcp"with"GroupDocs.Comparison.Mcp@26.7.4"inargs. Pinning is recommended for shared / committed configs to avoid surprise upgrades.
Usage with VS Code / GitHub Copilot
NuGet.org generates a ready-to-use mcp.json snippet on the package page.
Copy it directly into your .vscode/mcp.json.
Alternatively, add manually to .vscode/mcp.json:
{
"inputs": [
{
"type": "promptString",
"id": "storage_path",
"description": "Base folder for input and output files.",
"password": false
}
],
"servers": {
"groupdocs-comparison": {
"type": "stdio",
"command": "dnx",
"args": ["GroupDocs.Comparison.Mcp", "--yes"],
"env": {
"GROUPDOCS_MCP_STORAGE_PATH": "${input:storage_path}"
}
}
}
}
Same pinning rule as above — swap
"GroupDocs.Comparison.Mcp"for"GroupDocs.Comparison.Mcp@26.7.4"to lock to a specific release.
Usage with Docker Compose
cd docker
docker compose up
Edit docker/docker-compose.yml to point volumes at your local documents folder.
Documentation & guides
Step-by-step deployment guides and a published-package integration test suite live in the companion repo GroupDocs.Comparison.Mcp.Tests:
- Install from NuGet —
dnx, global tool, pinned vs always-latest - Run via Docker
- Verify on the MCP registry
- Use with Claude Desktop
- Use with VS Code / GitHub Copilot
- Run the integration tests
That repo also exercises every advertised tool against the published NuGet artifact on Linux, macOS, and Windows in CI — so the snippets above are verified end-to-end on every release.
License
MIT — see LICENSE
<!-- mcp-name: io.github.groupdocs-comparison/groupdocs-comparison-mcp -->
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.