Immich MCP Server: Give AI Agents Access to Your Photos, Albums, and ML Recognition Data

Why Give AI Access to Immich?

Immich is the self-hosted Google Photos alternative with ML-powered face recognition, object detection, smart search, and automatic organization. It handles photo and video backup from mobile devices, EXIF metadata, shared albums, and timeline views.

But when you're coding in Claude or Cursor, your AI assistant can't see your Immich library. It can't search your photos, check if the ML pipeline is running, or query album metadata. You end up opening the Immich web UI every time you need context about your photo library.

With DataFaucet, you browse your Immich dashboard once and it captures the API calls as MCP tools. Now your AI coding assistant can query photo data, check ML status, and browse albums directly.

What AI Agents Can Do With Immich Access

Once DataFaucet creates an MCP server from your Immich session, AI agents can:

• Search photos by date, location, person, or object tag
• Read EXIF metadata including camera model, GPS coordinates, and exposure settings
• Check ML pipeline status for face detection, object recognition, and CLIP embeddings
• Browse albums with photo counts, sharing status, and cover images
• View library statistics including total photos, videos, storage usage, and duplicates
• List recognized faces with assigned names and clustering groups
• Check server job queue for background tasks like thumbnail generation and video transcoding

Setting Up Immich as an MCP Server with DataFaucet

1. Open your Immich web interface and navigate through timeline, albums, search, and admin pages
2. DataFaucet captures the API calls your browser makes to Immich's backend
3. Click "Create MCP Server" and DataFaucet generates structured tools from those endpoints
4. Add the MCP server URL to your AI client (Claude, Cursor, Windsurf)
5. Your AI can now query Immich data in real-time

The entire process takes about two minutes. Works with any self-hosted Immich instance.

Example Queries Your AI Can Answer

After connecting, try asking your AI:

• "How many photos were taken in Paris last summer?"
• "Is the ML face recognition pipeline still processing?"
• "What's the total storage usage for my photo library?"
• "Show me albums shared with external users"
• "How many unassigned faces remain in the clustering queue?"
• "What camera models are most common in my library?"

Your AI can answer these instantly because it has real-time access to your Immich API through MCP tools.

DataFaucet vs Immich's REST API

Self-Hosted Immich Works Perfectly

Since DataFaucet captures API calls from your browser, it works with any Immich instance on your network. Your AI assistant gets access to the same data you see in the web UI, whether Immich runs on a home server, a NAS, or a cloud VM.

Use Cases

• Photo management scripting: AI reads library stats while you write organization scripts
• ML pipeline monitoring: Check if face clustering or object detection jobs are complete
• Storage planning: AI reports on library growth and storage consumption trends
• Album automation: AI reads album structure to help you build sharing workflows
• Metadata queries: Search photos by EXIF data without opening the Immich UI

Get Started

DataFaucet turns your Immich web interface into structured MCP tools in under two minutes. No API keys to generate, no custom code, no Docker configuration changes.

Create your Immich MCP server and give your AI coding assistant full visibility into your self-hosted photo library.