Self-hosted · AI-native · open source
Drop KenNook on a folder of photos and videos and search them like you’d search Google — “beach trips with the kids”, “the dog as a puppy” — privately, on your own machine.
No cloud. No uploads. No subscription.
Semantic search, face grouping, video transcripts, and a local LLM that answers questions about your library — no API keys, no internet required.
Runs entirely on your hardware. Zero telemetry, zero uploads. Your memories never leave the room.
Open source under AGPL-3.0. No lock-in, no subscription — free to self-host, forever.
Runs on your machine — no database server, no Python, no Docker. You’ll need Node 20+, pnpm, and ffmpeg.
# clone, install, index your media, then open it git clone https://github.com/kennook/kennook cd kennook && pnpm install pnpm indexer ~/Pictures pnpm dev # → http://localhost:3000
Full setup & prerequisites in the README.