Scope

Point it at a feed. Ask what's in frame.

Attach a public Maryland CHART camera or a still URL. Cloudflare grabs a frame, Workers AI reads the scene, and occupancy dwell says how long the location has been busy. This is free.

Open the booth

Endpoints

GET https://scope.nextautomatica.com/health
GET https://scope.nextautomatica.com/ui — search a camera and ask
GET https://scope.nextautomatica.com/v1/cameras?q= — public CHART cameras
POST https://scope.nextautomatica.com/v1/watch — attach a feed
POST https://scope.nextautomatica.com/v1/ask — grab a frame and answer
POST https://scope.nextautomatica.com/v1/frame — snapshot only
GET https://scope.nextautomatica.com/v1/status?feed_id=
GET https://scope.nextautomatica.com/f/{id} — last JPEG
https://scope.nextautomatica.com/mcp — Streamable HTTP MCP (cameras, watch, ask, frame, status)

What it will say

Anonymous vehicle and people counts
Where people stand in frame (clothing/pose, no faces)
How long the location has been occupied across samples
Not plates, tags, faces, or who someone is — especially not on public highway cameras