Give an AI agent its own key
Create a key for one AI agent, limited to the profiles, sites and actions it needs, see what it did, and revoke it at once.
Give every AI agent its own Browz key: limit it to the scopes, profiles and sites it needs, cap how many calls it can make a minute and a day, and revoke it, or stop every agent at once, whenever you want. Every call is logged on your own computer for 90 days, with the page address but not its query, and never the page content. These controls are free on every plan and run entirely on your computer.
You manage keys and stored sign-ins in the Browz app, under Automate (Agent keys and Saved sign-ins), or from the command line (browz agent-token and browz credential). The browz command is built into the agent, at Browz.app, Contents, Resources, browz-agent. On a server without the desktop app, run the same commands as browz-mcp agent-token.
- Create a key: browz agent-token create --name "shop checker" --profiles own --sites shop.example.com. The key, which starts with blcl_, is printed once. Browz keeps only its hash, so save it where only you and the agent can read it.
- Choose what it may do. By default a key may read and change profiles and drive their browser. --read-only allows reading only, and --scopes picks exactly. Running script, cookies and downloads need browz:browser:advanced, and stored sign-ins need browz:credentials:use; neither is in the default set.
- Limit where it may go. --profiles own allows only the profiles the key creates; --profiles with a list of profile ids, or --group with a group id, names them. --sites allows only those sites and their subdomains: a page off the list is refused, and after a redirect off it the agent cannot read or act on that tab until it goes back.
- Give the key to the agent, never the API token. For an MCP client, set BROWZ_LOCAL_AGENT_TOKEN, or BROWZ_LOCAL_AGENT_TOKEN_FILE naming a file that holds it, in the server's environment, not on its command line. For the local API and the SDKs, send it as the bearer token or set BROWZ_TOKEN to it.
- See what it did: browz agent-token log --agent followed by the key id. Each row has the time, the tool, the profile, the page address without its query, and the result. Page content, typed text and screenshots are never logged. Rows are kept 90 days.
- Stop it: browz agent-token revoke followed by the key id ends it for good; add --stop-browsers to stop the browsers it started. browz agent-token stop-all suspends every key at once and stops their browsers, and browz agent-token resume followed by a key id lets that one carry on.
A key lasts 30 days unless you choose --days, up to 90, or --no-expiry. It makes at most 60 calls a minute and 10,000 a day unless you change --per-minute and --per-day. At most 20 keys can be active at once. A key can never create or manage keys, or read the log.
A key governs what an agent does through Browz. Programs that run as the same computer user as Browz can read its data folder, so run an agent you do not trust under another user or in a container.
Let an agent sign in without seeing the password →
Still stuck? Export diagnostics and contact support.