Browser & data capture

Browser Use

browser-use/browser-use

The standard way to let an AI agent drive a real browser — open pages, click, type, and fill forms.

Repository
Stars
109k
Forks
11.9k
Open issues
342
Language
Python
Licence
MIT
Created
2024-10-31
Last push
2026-08-06
Effort to adopt

Some wiring

A few days of integration work — credentials, data mapping, a deploy target.

01

What it actually does

Browser Use makes websites usable by agents. It opens pages, clicks buttons, types, and fills forms the way a person does. You describe the task and it completes it.

The examples in the repository are unusually concrete: filling in a job application from a résumé, extracting structured data about followers and exporting it as CSV, QA automation across a web app. That is the honest shape of the value — the tedious web work nobody wants to do manually.

It is built on Playwright, so the underlying automation is battle-tested, with the agent layer handling the "figure out which button" problem that makes traditional selectors brittle.

108,000 stars, MIT licensed, with a cloud offering alongside the open-source library. It is the reference implementation in this category, which also means it is what other tools benchmark themselves against.

02

Who it's for

  • 01

    Growth and sales teams doing repetitive web research or data collection

  • 02

    QA teams who want tests that do not break on every selector change

  • 03

    Ops people automating a web workflow that has no API

  • 04

    Engineers who need browser automation without maintaining brittle scripts

SalesMarketingOperationsEngineeringSaaSE-commerceAgencyAny industry
03

Where it earns its keep

  • Extracting structured data from sites that offer no export
  • Filling long repetitive forms across many records
  • QA flows that survive UI changes better than selector-based tests
  • Competitive research across many sites on a schedule
  • Automating an internal tool that was never given an API
04

Use it, or skip it

Reach for it when

  • The target has no API and the work is genuinely repetitive
  • Selector-based automation keeps breaking on UI changes
  • You need headless automation running in CI or on a server
  • The task is describable in a sentence but tedious to do by hand

Skip it when

  • An API exists — use it, it will be faster, cheaper, and more reliable
  • The site's terms prohibit automated access; check before you build
  • You need logged-in sessions to carry cleanly, where a shared-browser tool fits better
  • Volume is high enough that per-page model cost becomes the dominant expense
05

10 automations

  1. 01Sales

    Lead list enrichment

    Visit each company site, pull the details your CRM is missing, and write them back as structured fields.

  2. 02Marketing

    Competitor pricing watch

    Check competitor pricing pages weekly, diff against last week, and post material changes to Slack.

  3. 03Marketing

    Directory submission

    Submit your product to relevant directories by filling each form from one canonical record.

  4. 04Engineering

    Regression QA suite

    Describe critical user journeys in plain language and run them nightly against staging.

  5. 05Finance

    Supplier portal extraction

    Pull invoices or order status from supplier portals that never built an API, on a schedule.

  6. 06Support

    Review monitoring

    Collect new reviews across platforms and route negative ones to support with the context attached.

  7. 07Sales

    Event and conference research

    Extract speaker and attendee lists from event sites to build a targeted outreach list.

  8. 08Operations

    Compliance evidence capture

    Screenshot and archive specific pages on a schedule to evidence what was published when.

  9. 09Marketing

    Job posting intelligence

    Track competitor hiring pages to infer product direction and headcount investment.

  10. 10Operations

    Bulk account maintenance

    Apply the same settings change across many accounts in a tool with no bulk edit feature.

Want one of these running by Friday?

LimeDock builds these as real workflows inside your stack — deployed to your cloud, wired into your Slack and CRM, with the code in your repo. You pay a build fee and your own API keys, nothing else.

Book a workflow call
07

Source

Repository stats were read from the GitHub API and reflect the last time we refreshed this entry. The editorial breakdown above is LimeDock’s own analysis — we are not affiliated with browser-use.

https://github.com/browser-use/browser-use