Edge Docs

Run the Edge Near the Rig

The supported path today is a Linux machine near the rig that publishes live video, keeps a rolling clip buffer, and reports health back to the platform.

Open Quickstart Back to Docs Home

Supported Stack

This is the path we actively run and support right now.

💻

Linux Host

A small Linux box near the rig is the current supported runtime. It keeps the video path and clip path physically close to the hardware.

🖥

Python Supervisor

The current edge runtime is the Python supervisor in edge/edge_supervisor.py, which watches camera health, handles clip jobs, and reports edge health.

📹

RTSP Camera Source

The camera stays local. The edge host reads RTSP from the local network and publishes the feed upstream rather than exposing the camera directly to the internet.

🎥

Docker + ffmpeg

The proven path today uses Docker for live publishing and ffmpeg/ffprobe locally for probing, segmenting, and assembling clips.

Sample Config

This is the shape of the current edge config file.

edge/chicken-blaster.sample.json
{
  "node_slug": "chicken-blaster",
  "base_url": "https://yourmove.live",
  "auth_token": "REPLACE_WITH_OWNER_JWT",
  "rtsp_url": "rtsp://admin:password@192.168.0.106:10554/tcp/av0_1",
  "container_name": "ym-edge-chicken-blaster",
  "probe_interval_seconds": 2,
  "restart_cooldown_seconds": 20,
  "clip_buffer_enabled": true,
  "clip_pre_roll_seconds": 3,
  "clip_post_roll_seconds": 6
}

The config ties the local runtime to one private node, one local camera source, and one authenticated operator context.

What Each Field Means

Key runtime inputs
node_slug              platform node identity
base_url               yourmove app base URL
auth_token             operator JWT for edge API calls
rtsp_url               local camera feed
container_name         local publisher container
clip_*                 rolling-buffer behavior

This is meant to be readable and editable by a builder. The edge machine should feel like part of the rig, not a black box.

Operational Expectations

What the edge layer is responsible for in a healthy system.

1

Publish Live Video

The edge host keeps the live camera feed flowing into the platform’s ingest path.

2

Protect Clip Capture

The rolling buffer lives locally so clips can capture the action moment without depending on cloud-start timing.

3

Report Health

The supervisor reports heartbeats, recorder state, publisher state, clip backlog, and recent clip results back to the app.

4

Recover Locally

If the local camera or publisher stalls, the edge host is the right place to probe, restart, and recover without touching the rest of the platform.