Browser Relay
Browser Relay lets FleetQ cloud experiments drive a browser running on your own machine. A small relay agent on your computer polls the platform for commands, executes them locally, and posts results back — so automations can use a real local browser session without exposing it to the internet.
How it works
- You create a relay token and run the relay agent on your machine.
- The agent long-polls the platform for queued commands.
- When a cloud experiment needs the browser, it enqueues a command; the agent receives it, runs it locally, and posts the result back.
Relay token
Create a relay token in Settings → Relay. The token is a 64-character cryptographically-random value, shown once at creation (only its SHA-256 hash is stored), scoped to your team, and valid for 90 days. The relay agent authenticates by sending the token as a bearer token on every request.
The relay agent
Download the relay-agent package directly from the platform:
curl -fsSL https://fleetq.net/get/relay-agent -o fleetq-relay-agent.zip
The download is a Node.js project (fleetq-relay-agent).
Unzip it, install dependencies, and run it with your relay token configured.
Command channel
The agent talks to the platform over two relay-token-authenticated endpoints:
| Method | Path | Purpose |
|---|---|---|
| GET | /api/relay/commands | Long-poll for the next command (blocks up to 25s, then returns a timeout to re-poll) |
| POST | /api/relay/result | Submit the result of an executed command (one POST per command) |
Each poll blocks for up to 25 seconds waiting for a queued command. A poll also updates the agent's
heartbeat, so the platform knows the relay is connected. When no command arrives within the window the
endpoint returns {"type":"timeout"} and the agent
re-polls immediately.