Add project: pyimgtag

[kurok]

Basic info

Application name: pyimgtag
Application repo link: https://github.com/kurok/pyimgtag
Application home link: https://pypi.org/project/pyimgtag/
Application description: Command-line photo tagger that uses a local vision model (Gemma via Ollama by default) to add searchable tags, scene categories, and EXIF-GPS-derived location to images, then lets you query, score, and clean up the photo library from the terminal. Apple Photos write-back on macOS. Local by default; optional remote or cloud backends (Anthropic/OpenAI/Gemini) via --backend.

Qualifications

• Free software with an online source repository.
• Using Python for a considerable part of their functionality.
• Well-known, or at least prominently used in an identifiable niche.
• Maintained or otherwise demonstrably still functional on relevant platforms.
• An application, not a library or framework.

Additional notability info

pyimgtag fills a practical niche for photographers managing large local libraries: it automates tagging using an on-device LLM (no data leaves the machine by default), enriches images with EXIF GPS-derived location names via Nominatim, and provides CLI subcommands to query, score, and purge low-quality results. Cross-platform (macOS/Linux/Windows); macOS users additionally get write-back into Apple Photos. Available on PyPI and configured via CLI flags, not code.

Entry added to projects.yaml under the graphics topic, tagged ai, console, linux, windows, mac.