zero-transcode-media-skilllisted
Install: claude install-skill 4ntiDandruff/zero-bloat-skills
# Zero-Transcode Edge Media Skill
Real-time CCTV and bench monitoring streaming pattern for low-power edge hardware (Armbian TV boxes, repurposed laptops, or edge single-board computers) without exhausting CPU cores on software video re-encoding.
---
## 1. The Cost of Transcoding on Constrained Hardware
- Re-encoding incoming H.264/H.265 video streams via FFmpeg CPU decoders absorbs 80-100% of CPU capacity on non-GPU edge appliances.
- High thermal throttling, frame stuttering, and eventual system freeze follow within hours.
---
## 2. Circuit Solution: Native RTP Pass-Through (go2rtc)
Utilize a compiled Go binary (`go2rtc`) that forwards raw RTP packets directly to browser endpoints without decoding video frames:
### `go2rtc.yaml` Configuration:
```yaml
streams:
bench_camera:
- rtsp://admin:pass@192.168.1.100:554/stream1
workshop_overhead:
- rtsp://admin:pass@192.168.1.101:554/live/ch0
api:
listen: ":1984"
webrtc:
listen: ":8555"
```
---
## 3. Data Flow & Efficiency Comparison
| Metric | FFmpeg Transcoder | go2rtc Pass-Through | Real Gain |
|---|---|---|---|
| CPU Utilization | 65% - 95% | 0.2% - 1.5% | ~98% CPU reduction |
| Memory Allocation | ~250 MB / stream | ~15 MB / stream | ~94% RAM reduction |
| Glass-to-Glass Latency | 3.0 - 5.0 seconds | <0.3 seconds (WebRTC) | Sub-second real-time |
---
## 4. Minimalist Web Client Integration
Render live video feeds into browser interfaces using standard HTML5 WebRTC:
```html
<div class="rounded-xl