MCP for AI Agents (Claude, ChatGPT & more)

Connecting an AI assistant to freedam means it can actually work in your library, not just talk about it. Instead of copy-pasting filenames into a chat window, you connect Claude, ChatGPT, or any assistant that supports the Model Context Protocol (MCP - the open standard for giving AI tools access to other software), and it can search, fetch, upload, organize, and share assets on your behalf.

What it does

freedam includes its own MCP connection out of the box. There is nothing extra to install or run - it is part of the product, available on your instance. Through it, an assistant can:

• Search the library using natural language, the same smart search people use
• Look up an asset with its metadata, previews, and usage rights
• Edit metadata - titles, descriptions, keywords, and structured fields, with the same validation as everywhere else
• Browse and create collections, including nested ones, and file assets into them
• Upload files, optionally straight into a chosen collection
• Create share links with expiration dates and access controls
• Use your approved vocabularies, so it picks valid terms instead of inventing its own
• Run bulk changes across many assets at once and report back on their progress

That covers real workflows: an assistant can go through a fresh upload batch, fill in missing metadata using your vocabularies, sort assets into collections, and hand back a share link when it is done. There are no special AI-only powers - everything an assistant can do is something the product already allows.

How it works

You create an API token in Settings → API Tokens, choosing exactly what the assistant is allowed to do, then paste your instance address and the token into your assistant's settings:

{
    "mcpServers": {
        "freedam": {
            "url": "https://your-instance.example.com/mcp/freedam",
            "headers": { "Authorization": "Bearer YOUR_TOKEN" }
        }
    }
}

The important part: the assistant goes through the same permission checks as a human user. A token that can only read the library cannot be talked into deleting an asset, no matter how creative the prompt. And every action lands in the same audit trail as everything else, so you can always see which token did what, when, and to which asset.

Good to know

• Permissions are enforced on every action. A token can never carry more power than the person who created it, and a missing permission means a refused action, not a quiet workaround.
• Speed limits apply. An assistant works within the same per-token limits as any other integration, so a runaway loop hits a ceiling instead of your library.
• Tokens are shown once. Copy the token when you create it; freedam only keeps a fingerprint. If you lose it, revoke it and make a new one.
• Self-hosted instances get the same connection - it is part of the product, not a hosted add-on.

The MCP connection sits on the same foundation as the REST API, so anything an assistant starts, a script can finish. Want to see it drive a real library? Try the demo.