Developer platform · physical AI
Video, audio and sensor streams with your models, devices and people in one session. In our cloud or on your edge.
Two-way sessions, latency you can act on, and a catalog of models optimized for live video.
Video, results and control flow in both directions, no reconnecting.
Overlays, results and control come back inside a frame budget, not after a round trip to an inference API. The media and the model share a GPU, so there is no hop between seeing and deciding.
Models tuned and calibrated for live video. Bring your own weights: coming.
Every source publishes into the session. Models are participants too: a vision model, a robot policy or a generated world. Overlays, actions and control come back out live, and a person can join at any moment.
Built for developers
Create a session from your backend, invite your sources and your model, and subscribe to what comes back. No media pipeline to build, no GPU fleet to schedule, no backend for the first moving picture.
From your backend, with the model you want in it: one from the catalog.
Cameras over RTSP, drones over WHIP, robots over ROS 2, phones and browsers over the SDK. People, when they need to see.
Processed video, structured results per frame, alerts and control, in both directions, live.
// Backend: create the session with the model in it, // attach this device, hand the page the allocation: // rgb.sessions.allocate({ serviceID, ... }) import { InferenceClient } from "@rgb/browser"; const allocation = await fetch("/connection", { method: "POST" }).then(r => r.json()); const session = new InferenceClient({ origin }) .connection(allocation, { receiveVideo: true, onVideo: (stream) => { output.srcObject = stream; }, onResult: (result) => draw(result), // per frame }); await session.publishCamera(inputVideo); // camera in
See, act, learn. The three bars in the mark. One session, one clock and one SDK carry cameras, drones and robots, so what you build for fixed cameras today runs on a moving one tomorrow.
Every camera, drone and phone publishes into one session. Your model watches all of them and sends back overlays, depth and alerts. A supervisor joins from a browser, sees the alert beside the live view, and clears or escalates it.
Control rides the same session as the video, with takeover and handback built in. Operate drones, arms and vehicles remotely while the model and the operator see the same frame.
Every session is recorded on one clock, replayable against a new model and exported as episodes. Every takeover becomes a label, captured as a by-product of running the session.
Send recorded video into a session: field footage, archives, whole datasets. The same models run on it and return labels, depth and poses per frame with the file's own timestamps, ready to go into a training set. No need to jerry-rig GStreamer pipelines.
Per stream per month, cameras or robots, model included. Recording as an add-on.
Stop stitching cameras to inference APIs to dashboards to pagers. Put the model and your people in the same room as the stream.
A supervisor joins from a browser, sees the alert beside the live view, and clears or escalates. Models triage. People decide.
Early access · 2026
We are onboarding a small number of teams this quarter. Bring the cameras, drones or robots you already have; we run the session and you watch the number every week.