browser-use: an agent that drives the browser
An agent that drives the browser on its own — local model, no API key
- Video memory
- 12 GB
- Setup
- ~25 min
- Access
- port 7788
- Billing
- hourly, in BRL
You write the task in plain language and the agent really browses: clicks, types, scrolls, decides the next step from what it sees. This template brings up a language model on the machine itself, so no third-party API key is needed — and no login, visited site or collected data leaves the box.
Describe the task and the agent browses: fills forms, compares prices, walks catalogs, builds carts. Ships with a local LLM, so it works without any third-party API key, and nothing leaves the machine. Live browser view included.
What it is for
- Purchase automation: walk a catalog, compare prices, build a cart
- Fill repetitive forms in systems that have no API
- Collect data from pages whose layout changes every week
- Live browser view to audit every click the agent makes
How to deploy
- Create your account and add balance (card or Pix, no subscription).
- In the console, pick the browser-use template and a machine — the console hides the ones that do not meet the requirement.
- In about 25 minutes the setup finishes and the access address shows up in the panel, on port 7788.
Done? Just destroy the machine and billing stops with it. No contract, no minimum commitment.
FAQ
Do I need a paid API key?
No. The template runs a model on the machine itself and the interface already points at it. If you prefer an external model, you can configure that on screen.
Can I watch what the agent is doing?
Yes, there is a live browser view on another port, with a password generated during setup.
Does it work on sites with captcha?
Captcha and anti-bot protection are still a barrier. The agent is good at ordinary navigation flows, not at defeating protection.