Quick start

Note
These steps install the .NET build of the server — the only platform available today. Java, Python, and Node.js builds are planned; each will get its own install section under its platform. Platform-independent material (tools, use cases, licensing) lives here and applies to all of them.

Install and register the GroupDocs.Metadata MCP server in Claude Desktop, VS Code, Cursor, or any MCP client in three commands with the guided installer — then verify the setup automatically:

git clone https://github.com/groupdocs/GroupDocs.Mcp.Installer.git
cd GroupDocs.Mcp.Installer

# wizard: products, channel, clients, folders, license
./install-groupdocs-mcp.ps1 -Interactive
# preview - prints everything, changes nothing
./install-groupdocs-mcp.ps1 -DryRun
# apply + warm caches + verify the setup in one go
./install-groupdocs-mcp.ps1 -Verify

Then restart your AI client (Claude Desktop, VS Code, Cursor, …) so it picks up the new server.

Your first prompt

Put any document or image (say report.pdf) into the storage folder you chose, then ask your agent:

What metadata does report.pdf carry, and who is listed as the author?

The agent calls read_metadata and reports the properties. Nothing is uploaded — the file is read in place by the local server process.

Where to go next

  • Fresh machine? Follow your OS page — it includes the prerequisite bootstrap: Windows · Linux · macOS
  • Manual or single-client install (exact JSON per client): Register in AI clients
  • Docker-only shop: run ./install-groupdocs-mcp.ps1 -EmitCompose -Clients @() to generate a docker-compose.yml instead of client registration — see Configuration.
  • Licensing: the server works in evaluation mode out of the box; your existing GroupDocs.Metadata license applies — Licensing.