40 lines
1.3 KiB
Markdown
40 lines
1.3 KiB
Markdown
# visionLabs Frigate setup
|
|
|
|
This project now assumes Frigate is the backend that owns camera auth, camera discovery, and stream handling.
|
|
|
|
## Starter inventory
|
|
|
|
The first four cameras are currently modeled as:
|
|
|
|
- `cam195` -> `192.168.0.195`
|
|
- `cam196` -> `192.168.0.196`
|
|
- `cam201` -> `192.168.0.201`
|
|
- `cam228` -> `192.168.0.228`
|
|
|
|
All four are ONVIF-capable and should be added to Frigate using matching camera IDs where possible.
|
|
|
|
## Important rule
|
|
|
|
Do not commit raw camera passwords to this repo.
|
|
|
|
Keep camera credentials inside:
|
|
|
|
- Frigate config
|
|
- go2rtc config
|
|
- a local secret file outside version control
|
|
|
|
## What to update next
|
|
|
|
1. Set the real Frigate base URL in `config/cameras.json`
|
|
2. Confirm the Frigate camera IDs match `cam195`, `cam196`, `cam201`, and `cam228`
|
|
3. Verify that each camera snapshot is reachable at `/api/<camera>/latest.jpg`
|
|
4. Add snapshot tile loading in the SceneGraph UI
|
|
5. Decide which direct live endpoint format the Roku app should consume after snapshots are working
|
|
|
|
## Why this shape works
|
|
|
|
- Roku does not want to manage ONVIF directly
|
|
- Roku is happier when Frigate handles the camera-side protocol mess
|
|
- This keeps the app focused on layout, browsing, and playback
|
|
- Snapshot-first is the lowest-risk way to prove the app shape before live playback is finalized
|