Browser Use
Lets an agent run real browser-automation tasks on Browser Use Cloud - it drives a remote Chromium with an LLM, and you can watch it live and take over (for logins, 2FA, CAPTCHA) in an embedded browser window. Bring your own Browser Use API key; usage is billed to your Browser Use account.
What it can do
- run_browser_task - start a task ("log into X and download the latest invoice", "fill out this form", "compare prices across these sites"); returns a session, a live-view URL and a clean, embed-ready live view.
- wait_for_browser_task - follow a running task (short waits; re-poll for long ones).
- get_browser_task - check status / result.
- get_browser_recording - after a task finishes, save a durable demo (MP4 + final screenshot + downloaded output files) into your agent files.
- share_browser_session - get a clean, shareable live link for the running session.
- send_task - a follow-up step on a kept-alive session (same logged-in browser).
- stop_browser_task - stop / cancel a session.
Results come back as text (or typed JSON if you ask for a schema), plus screenshots and an optional recording.
Record demos
Ask the agent to record a task and you get a clean MP4 of the whole session - handy for demos and walkthroughs.
- Turn on recording. Set
enable_recordingand you get an MP4 covering the entire run. - Durable MP4 in your agent files. The raw
recording_urlis a short-lived presigned link (~1 hour). For a non-sensitive recorded session the MP4 is auto-saved into your agent files (it shows up in the files panel). You can also call get_browser_recording explicitly - it downloads the MP4 (plus the final screenshot and any downloaded output files) and saves them as permanent agent files -recording_file.public_urlis the link to share for a lasting demo. - Clean live view.
live_url_cleanis the same live session with the browser chrome (URL bar, tabs) hidden (ui=false), so an iframe embed or a screen recording shows only the page. Pick a theme withlive_view_theme(lightordark). The rawlive_url(with chrome) is still returned for taking over. - Shareable link.
share_url(or share_browser_session) is a clean link to the live session so someone else can watch it run. - Output files. Files the agent downloads during the task are listed and saved into
your agent files as well (
output_files), so a demo can ship its artifacts.
Keep a session open for manual testing
To keep a session open after the task finishes - for example to manually test a chat or
take over via the live view - set keep_alive=true. The session then stays alive so you
can interact via live_url (or send a follow-up step with send_task) instead of closing
right after the task ends.
A kept-alive session keeps reserving the browser and costs $0.02/hour while open, so always call stop_browser_task when you are done - don't leave an idle session running.
Limitations (cloud)
These are constraints of Browser Use Cloud (not FlyMyAI):
- No GIF export. Cloud produces an MP4 only; GIF is a self-hosted-only feature.
- No persistent public replay URL. Browser Use Cloud has no separate shareable replay
page -
share_urlpoints at the live session and only works while it is alive. For a permanent shareable demo, share the savedrecording_file.public_urlinstead. - Highlight overlays can't be turned off on cloud. The colored element-highlight boxes
the agent draws are always on; to record without them you would need to self-host Browser
Use (
highlight_elements=False). - Recording is whole-task. There is no manual start/stop - the MP4 covers the entire session.
- Presigned URLs expire fast.
recording_url~1 hour,screenshot_url~5 minutes, downloaded-file URLs ~15 minutes - which is exactly why get_browser_recording saves durable copies into your agent files. - Sensitive sessions. If a session used a saved login or
secrets, the recording / screenshot may show logged-in content, so the agent does not auto-save them - it only persists media when you explicitly ask for the demo. Share such recordings carefully.
Logins, 2FA and your passwords
The agent never needs your password in the chat. Two safe ways to log in:
- Log in once, in the live window (recommended). Start a task, open the live-view URL, and sign in yourself in the real remote browser. The session is saved as a Browser Use profile and reused next time - no password ever touches FlyMyAI or the model.
- Domain-scoped secrets. For automated logins, the credential is typed into the page by Browser Use (the model only ever sees a placeholder) and is scoped to the site's domain. It is never stored by us and never shown to the model.
When a task hits 2FA / SMS / CAPTCHA / "verify it's you" / a forced password change, the agent pauses and hands you the live view to complete it - then carries on.
How to get credentials
- Sign up at browser-use.com and open cloud.browser-use.com -> Settings -> API keys.
- Create a key (it starts with
bu_). - Paste it into the Browser Use MCP setup in FlyMyAI.
Fields to fill in FlyMyAI
| Field | Required | Notes |
|---|---|---|
BROWSER_USE_API_KEY | yes | Your bu_... key. Usage is billed to your own Browser Use account. |
Good to know
- Cost is metered on your Browser Use account (browser time, optional proxy, and the LLM). Each task runs under a budget cap, and the agent tells you the expected cost.
- Confirm before irreversible actions - the agent is told to ask before purchases, sends, or anything it cannot undo.
- Scope it. Tasks restrict the agent to the domains you allow; required when you pass a login.
- Pick a model per task - the default is a fast, cheap one; heavier sites can use a stronger model.
Troubleshooting
- "BROWSER_USE_API_KEY is not set" - paste your
bu_key in the tool setup. - Stuck on a login / verification - open the live-view URL and complete it yourself; the agent resumes and saves the login for next time.
- "Too many concurrent sessions" - your Browser Use plan caps concurrency; finish or stop running sessions, or upgrade your plan.