sentrysearch

Last week, I posted SentrySearch, a CLI for semantic video search using Gemini's embedding API. The #1 request was local model support. Turns out Qwen3-VL-Embedding can natively embed video into the same kind of vector space, no API, fully offline. Runs on Apple Silicon (MPS) and NVIDIA GPUs (CUDA). The 8B model needs ~18GB RAM, or use the 2B model on smaller machines. sentrysearch index /path --backend local Also added: similarity threshold to suppress weak matches, and a Tesla metadata overlay that renders speed/location onto matched clips. Details on the README.

The README on the GitHub has a section on this[0]: >Indexing 1 hour of footage costs ~$2.84 with Gemini's embedding API (default settings: 30s chunks, 5s overlap): >1 hour = 3,600 seconds of video = 3,600 frames processed by the model. 3,600 frames × $0.00079 = ~$2.84/hr >The Gemini API natively extracts and tokenizes exactly 1 frame per second from uploaded video, regardless of the file's actual frame rate. The preprocessing step (which downscales chunks to 480p at 5fps via ffmpeg) is a local/bandwidth optimization — it keeps payload sizes small so API requests are fast and don't timeout — but does not change the number of frames the API processes. [0] https://github.com/ssrajadh/sentrysearch#cost

Gemini Embedding 2 can project raw video directly into a 768-dimensional vector space alongside text. No transcription, no frame captioning, no intermediate text. A query like "green car cutting me off" is directly comparable to a 30-second video clip at the vector level. I used this to build a CLI that indexes hours of footage into ChromaDB, then searches it with natural language and auto-trims the matching clip. Demo video on the GitHub README. Indexing costs ~$2.50/hr of footage. Still-frame detection skips idle chunks, so security camera / sentry mode footage is much cheaper.

I built a semantic search CLI for dashcam footage using Gemini Embedding 2's native video embedding. The interesting part: Gemini Embedding 2 projects raw mp4 video directly into the same vector space as text, no captioning or transcription pipeline. You embed 30-second video chunks as RETRIEVAL_DOCUMENT, embed a text query as RETRIEVAL_QUERY, and cosine similarity just works across modalities. The tool splits footage into overlapping chunks, indexes them in a local ChromaDB instance, and auto-trims the top match from the original file via ffmpeg. Cost is about $2.50/hr of footage to index, queries are negligible. Definitely room to optimize: skipping still frames, scene detection for smarter chunking, etc.