Skip to content
fastbrowsefastbrowse

docs

Use fastbrowse.

Install fastbrowse, choose a browser and run a task. Each guide is also available as Markdown. For agents, use llms.txt or read llms-full.txt.

Source commit: 4372bfd6bb8ff39f7b14b1cbb7dba2285b9761d5.

back to fastbrowse.ai

Start

First run

Install fastbrowse with uv, choose a cloud or local browser, and run a first task from the command line.

Markdown·source

Use fastbrowse with your agent

Install the fastbrowse skill, configure a browser and delegate cited website tasks with explicit authorization.

Markdown·source

Browsers and credentials

Signed-in sites and secrets

Reuse a signed-in Chrome profile, run a cloud profile, or type origin-scoped secrets from a Bitwarden vault.

Markdown·source

Attach to a running browser

Drive a Chrome or Electron window over its DevTools port, or attach to a window that is already open.

Markdown·source

Models and run statuses

Models

Choose planning and reading models, override them for specific tasks, and set reasoning effort.

Markdown·source

Run statuses and exits

Every status and exit code, what complete means, and how a run reports an exhausted budget.

Markdown·source

Libraries and tools

Use it from Python

Embed run_task, request schema-validated data, read citations and events, stream frames, and attach over CDP.

Markdown·source

Use it from JavaScript

Install the TypeScript SDK, run a task, and keep secrets, authorization and output validation in your process.

Markdown·source

Use it from an MCP client

Serve the browse tool over MCP, with server-side limits, fixed browser choices and HTTP authorization.

Markdown·source

Trust

Safety model

How fastbrowse checks authorization, handles secrets and treats page content as data.

Markdown·source

How it works

The step loop, reading and citations, prompt limits, run events and images, and browser capabilities over CDP.

Markdown·source