Start a private help session.
Create a temporary QR/code. Nothing is shared until one helper joins and you explicitly choose a screen or window.
One person shares exactly the screen or window they choose. One helper watches through encrypted WebRTC, can speak or send live guidance, and never receives arbitrary control of the computer.
Start ASSIST →Create a temporary QR/code. Nothing is shared until one helper joins and you explicitly choose a screen or window.
Joining does not request camera, microphone or screen access. You receive only what the host explicitly shares.
ASSIST cannot choose a screen silently. The host selects a tab, window or display through the browser's native capture picker.
v1 has no keyboard injection, mouse control, clipboard access, shell execution or arbitrary native commands.
Screen and voice tracks travel peer-to-peer when possible. TURN may relay encrypted packets transiently when direct NAT traversal fails.
The signaling service stores short-lived pairing and SDP/ICE state only. Chat and guidance use the encrypted DataChannel.
Both participants use the website. Pairing is a short-lived code or QR and one helper is admitted per session.
ASSIST cannot silently capture a display. Screen sharing starts only from the browser's native picker after an explicit host action.
Media travels over WebRTC. TURN can relay encrypted packets when required, while the 2XBR signaling endpoint keeps only bounded session and SDP/ICE state.
No keyboard injection, mouse control, clipboard extraction, shell execution, file browsing or hidden persistence is exposed by ASSIST.
Most support sessions start with one need: let another person see what you see and explain the next move. ASSIST v1 solves that layer first, so permission and risk stay understandable.