SentrySearch
- 1Runs onLinux, Mac, self-hosted, Windows
- 2Visual searchYes
- 3Audio searchNo
- 4Multimodal searchYes
- 5API accessNo
Overview
SentrySearch searches video footage by text or reference image and can return trimmed matching clips. It splits videos into overlapping chunks, embeds footage and queries into a shared vector space, and stores video vectors in a local ChromaDB database. Search options include text, image, anomaly highlights, and optional reranking of candidate clips. The project documents embedding backends including Gemini, Alibaba DashScope Qwen Cloud, LiteLLM, local Qwen3-VL, and MLX for Apple Silicon. Its local backend runs on the user's machine without an API key and is described as private and offline-capable. With Qwen Cloud, local video chunks are uploaded by the official SDK to DashScope-managed temporary object storage for processing. The scanner recursively finds MP4 and MOV files, including footage not from Tesla Sentry Mode. An optional Tesla overlay can use supported footage metadata for speed, GPS, and time information. The project requires Python 3.11 or later and FFmpeg or imageio-ffmpeg. It is open source under Apache-2.0 and supports Linux, macOS, Windows, and self-hosting.
Who it is for
SentrySearch may suit people who need to find moments in video using text or images, including users working with Tesla footage. Its local backend is relevant to those seeking machine-side processing without an API key.
What is good
- Searches by text and reference images.
- Can rank anomalous clips and trim them.
- Local backend runs without an API key.
- Scans MP4 and MOV files recursively.
- Apache-2.0 license.
What to know first
- Requires Python 3.11 or later and FFmpeg or imageio-ffmpeg.
- Local inference requires CUDA or Apple Metal.
- Still-frame detection can miss subtle motion.
- Events across chunk boundaries may not match perfectly.
Verdict
SentrySearch offers several search backends and a local processing option, with clear runtime requirements. Account for its motion-detection and chunk-boundary limits when searching for brief or subtle events.
Compared on AI video search tools
- Visual search
- Yesgithub.com
- Audio search
- Nogithub.com
- Multimodal search
- Yesgithub.com
- API access
- Nogithub.com
Facts
- Purpose
- SentrySearch performs semantic search over video footage and returns trimmed clips matching a search.github.com · 7 Oct 2026
- How it works
- It embeds video chunks and text or image queries into a shared vector space, stores video vectors in a local ChromaDB database, and matches queries against them.github.com · 7 Oct 2026
- Search modes
- It supports text search, image search, anomaly highlights, and optional reranking of candidate clips.github.com · 7 Oct 2026
- Embedding backends
- The project documents Gemini, Alibaba DashScope Qwen Cloud, LiteLLM, local Qwen3-VL, and an MLX backend for Apple Silicon.github.com · 7 Oct 2026
- Privacy option
- The local backend runs on the user's machine without an API key, which the README describes as private and offline-capable.github.com · 7 Oct 2026
- Cloud data handling
- For Qwen Cloud, local video chunks are uploaded to DashScope-managed temporary object storage by the official SDK before the API processes them.github.com · 7 Oct 2026
- Integrations
- The README documents LiteLLM gateways and handoffs to the maker's SentryMerge and SentryBlur sibling tools.github.com · 7 Oct 2026
- Tesla support
- An optional overlay extracts speed, GPS, and time metadata from supported Tesla driving footage and can add location labels through optional OpenStreetMap reverse geocoding.github.com · 7 Oct 2026
- Supported footage
- The directory scanner recursively finds MP4 and MOV files, including footage that is not from Tesla Sentry Mode.github.com · 7 Oct 2026
- Requirements
- The README lists Python 3.11 or later and FFmpeg or its bundled imageio-ffmpeg package; local inference requires CUDA or Apple Metal, and its macOS video decoder requires system FFmpeg.github.com · 7 Oct 2026
- Limits
- Still-frame detection is heuristic and can miss subtle motion, while events spanning chunk boundaries may not be matched perfectly.github.com · 7 Oct 2026
- Usage costs
- The README estimates Gemini indexing at about $2.84 per hour of footage with its default settings and says local-backend calls use no API quota.github.com · 7 Oct 2026
- License
- The public GitHub repository identifies its license as Apache-2.0.github.com · 7 Oct 2026
- Search methods
- Users can search with text queries or reference images.github.com · 8 Oct 2026
- Local processing
- The local backend runs without an API key and processes footage on the user's machine.github.com · 8 Oct 2026
- Video handling
- It splits videos into overlapping chunks, stores embeddings in a local ChromaDB database, and can trim matching clips.github.com · 8 Oct 2026
- Highlights
- The highlights command ranks anomalous clips in an index and can trim them automatically.github.com · 8 Oct 2026
- Tesla overlay
- An optional overlay can display Tesla dashcam speed, date, time, city, and road name when supported metadata is available.github.com · 8 Oct 2026
- Privacy
- The README describes the local backend as private and says it runs entirely on the user's machine.github.com · 8 Oct 2026
- Compatibility
- The directory scanner recursively finds MP4 and MOV footage, including footage not recorded in Tesla Sentry Mode.github.com · 8 Oct 2026
- Maker
- The maintainer's GitHub profile names Soham Rajadhyaksha and lists Fremont, California.github.com · 8 Oct 2026
Best SentrySearch alternatives
See all 20
Where it ranks on Specifiction
- Best AI Video Search Tools in 2026#10 of 24
Is SentrySearch yours?
Claim it for free: prove the domain, then correct facts, plans and screenshots. An editor reviews every change.
Sources
- github.com/ssrajadh/sentrysearch· checked 7 Oct 2026
- github.com/ssrajadh· checked 8 Oct 2026