Stars
Forks
Watchers
Developer links
OpenNVR
OpenNVR turns your own cameras and hardware into a video platform where recording, live viewing, and AI analysis all stay inside your perimeter. Any ONVIF or RTSP camera connects, including an Android phone running the companion camera app, and footage is written in one-minute chunks whether or not any model is running. Detection overlays appear moments after a camera is added, and you search recorded events in natural language instead of scrubbing timelines. Twelve installable applications cover intrusion detection, loitering, occupancy counting, line crossing, license plates, camera tampering, gate watching, and a doorbell that recognizes your family, each declaring its camera and network permissions before an administrator approves it. Swapping the underlying model changes nothing else, because every detector, plate reader, face recognizer, or vision-language model answers through the same adapter contract. Two Python SDKs let you write your own app as a rule over platform events or wrap your own model as a first-class capability. Cameras, zones, and apps surface in Home Assistant automatically through MQTT discovery. Offline and local-only modes return errors to every cloud route until an operator opts in, model weights are fingerprinted with SHA-256 and re-checked every minute, and an append-only audit log traces each alert to the exact weights that produced it. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.
Benefits
- Any Camera, Any Vendor
- ONVIF and RTSP cameras from any manufacturer connect alongside Android phones and webcams, with footage recorded in one-minute chunks that never depend on AI services being available.
- Models Swap Like Software
- Object detection, plate reading, faces, pose, and vision-language models all answer through one adapter contract, so upgrading weights or switching detectors changes nothing else in the system.
- Proof for Every Alert
- Model weights are SHA-256 fingerprinted and re-checked every minute, and one correlation ID traces each alert through the gateway to the exact weights that produced it.
- Offline by Default Design
- Default-deny offline and local-only gates block cloud routes and non-local models until an operator opts in, and the camera network can sit behind a second NIC with no route out.
Features
- Twelve Installable Apps
- Intrusion, loitering, occupancy, line crossing, plate recognition, tamper detection, gate watch, and family-aware doorbell apps install in one click with declared permissions.
- Natural-Language Search
- Every event lands in one store with its best frame, model claims, and confidence, so the search page, the agent, and installed apps all answer from identical evidence.
- Two Python SDKs
- The app SDK turns your rule into a platform application over its events, while the adapter SDK wraps your own model behind the published capability contract.
- Home Assistant Discovery
- Cameras, zones, and apps appear as devices in Home Assistant automatically through MQTT discovery, with nothing extra to install on the Home Assistant side.
- Tiered Detection Pipeline
- A cheap always-on Tier-0 detector scans substreams at low frame rates and wakes expensive models only when footage deserves a closer look, saving substantial compute.