Labsco
MCP SERVER

Smart Photo Journal MCP Server

by Siddhant-K-code

Search your macOS Photos library by place, label or person — with fuzzy matching — and get a read on when you actually take photos.

Image Analysis, OCR & Vision
Summary
The searches Photos already supports, available to an assistant.

Nothing here uploads anything — it reads the local Photos library after macOS grants permission. The analysis tool is the one that does something the Photos app does not put in front of you: aggregate patterns over time rather than a search result. Names and places are matched loosely, which is the right default when you half-remember what an album was called.

What it is

A server over the macOS Photos library. It answers the questions a folder of images cannot: photos from a place, photos with a person in them, photos under a label, and the patterns across all of them.

What you get
  • location-search finds photos taken in a named place and reports how many it found
  • label-search finds photos by keyword or label such as Birthday, Beach or Vacation
  • people-search finds photos containing a named person
  • photo-analysis reports patterns across the library, including the most common times and days you take photos
  • Fuzzy matching across those searches, so an approximate name still returns results
Requirements

macOS with a Photos library loaded, plus uv. Clone the repository, run `uv sync`, and launch with `uv run server.py` or let the client start it. macOS will ask you to authorise access to your photos the first time it runs; the library is read locally.

Setup effort

One command — uvx memory-journal-mcp-server