Browser Use
Also known as: browser-use, Browser Use Cloud
Open-source library and cloud platform enabling AI agents to browse and interact with web pages autonomously.
Browser Use is an open-source library that lets AI agents drive a real web browser. Its core idea is to replace brittle automation scripts, the CSS selectors and XPath that break the moment a site changes its markup, with natural language: a developer describes a task in plain English, and Browser Use figures out how to carry it out against whatever page it encounters, navigating, clicking, filling forms, and extracting data, even on sites it has never seen before. The project has become one of the most widely adopted open-source building blocks for browser agents.
The technical heart of the library is how it presents a web page to a language model. Browser Use distills a messy, sprawling DOM down to the elements that actually matter, labeling the interactive components and stripping the noise, which both makes the model's decisions more reliable and cuts the number of tokens each step consumes. Because the agent reasons from the page's current visual and structural state rather than hardcoded selectors, its automations self-heal when a layout changes, which is what makes Browser Use practical for legacy web apps, government portals, and the many sites that have no usable API.
Browser Use is deliberately model-agnostic. It works with frontier models from several providers as well as local models, so teams keep full control over model choice and cost, and it ships an MCP server so assistants like Claude Code and Cursor can invoke browser actions, navigating, filling forms, extracting data, and managing tabs, as native tools. Built on top of established browser automation under the hood, it offers a Python and TypeScript SDK, a command-line interface, and a desktop app.
The open-source library is free under a permissive license and can be self-hosted, while a companion cloud service handles the harder operational parts of running agents at scale: large numbers of parallel sessions, stealthy undetectable browsers, CAPTCHA handling, and residential proxies across many countries. In the typical production stack, Browser Use supplies the agent's reasoning and web-interaction logic, paired with managed browser infrastructure and a model of the team's choosing.
Vendor details
Canonical URL
https://browser-use.com/
Category
Browser / computer-use agent
Company status
independent
Use cases & customers
Target customers
Deployment options
In practice
Your scraping and automation scripts break every time a target site changes its layout. Browser Use lets your agent reason from the live page instead of hardcoded selectors, so the automation self-heals across redesigns.
You want to automate a site that has no API and sits behind a normal web UI. Browser Use takes a plain-English task and navigates, clicks, fills forms, and extracts data against the real page, model of your choice.
You want Claude Code or Cursor to actually operate websites, not just describe steps. Browser Use ships an MCP server that exposes navigation, form-filling, data extraction, and tab management as native tools to your coding agent.
Agentic Index coverage score
11.0 / 14 capabilities · 79%
| Integrations & Tool CallingAgentic AI Index Expansion Evidence Report | Full |
|---|---|
| Workflow OrchestrationAgentic AI Index Expansion Evidence Report | Full |
| Knowledge Grounding & RAGAgentic AI Index Expansion Evidence Report | Unable to verify |
| Human Oversight & GuardrailsAgentic AI Index Expansion Evidence Report | Full |
| Security, Identity & GovernanceAgentic AI Index Expansion Evidence Report | Partial |
| Observability & AuditabilityAgentic AI Index Expansion Evidence Report | Full |
| Memory & State PersistenceAgentic AI Index Expansion Evidence Report | Partial |
| Deployment & Data ResidencyAgentic AI Index Expansion Evidence Report | Full |
| Prebuilt Agents, Templates & PacksAgentic AI Index Expansion Evidence Report | Full |
| Triggers & Channel CoverageAgentic AI Index Expansion Evidence Report | Partial |
| Model Flexibility & RoutingAgentic AI Index Expansion Evidence Report | Full |
| APIs, SDKs & MCP ExtensibilityAgentic AI Index Expansion Evidence Report | Full |
| Testing, Debugging & OptimizationAgentic AI Index Expansion Evidence Report | Partial |
| Browser & Computer UseAgentic AI Index Expansion Evidence Report | Full |
The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Recent platform changes
Browser Use updated its Browser Harness v4 API to support client-controlled session recording, allowing API callers to explicitly enable or disable recording via the browserSettings.record parameter. The release also added a new v3:browsers:read OAuth scope that grants external applications read access to existing browser sessions.
Bears on: Observability / auditability
View sourceBrowser Use released the Browser Harness agent API to the public. The update also introduced new AI models across the platform and lowered the pay-as-you-go top-up minimum.
Bears on: MCP / tool calling / API
View sourcebrowser-use released 0.13.1, adding support for Claude Fable 5 and changing Anthropic behavior so tool choice is auto when using thinking. This is a dated compatibility release; current Fable 5 availability has not been separately verified.
Bears on: MCP / tool calling / API
View sourcePricing
Usage-based (~$0.06/hr sessions, $10/GB proxy) · OSS free · $10 credits
usage
Included quota
Open-source Browser Harness library: free (self-run, pay only your own LLM tokens). Cloud signup: $10 free credits (no permanent free tier). Pay-as-you-go: $0.06/hr sessions, $10/GB proxy, token 1.2x (or BYO key + 0.2x). Starter ($100/mo): $100 included credits, $7.50/GB proxy, up to 50 concurrent sessions. Business ($500/mo, $400 annual): $0.03/hr sessions, $5/GB proxy, up to 250 concurrent sessions. Scaleup (~$2,500/mo): higher concurrency + 60% off proxy. Enterprise (custom): ZDR/no-training/HIPAA/BAA, dedicated browser pool, on-prem, custom SLAs. SOC 2 certified; stealth browsers, captcha bypass, Python/TypeScript SDKs.
What is public
Browser Use (browser-use.com - AI browser-automation; OSS library 'Browser Harness' with 79,000+ GitHub stars and 1.3M+ monthly downloads, plus a fully-hosted cloud agent) is USAGE-BASED, not a flat monthly subscription. New accounts get $10 in free signup credits (no permanent free cloud tier; the open-source library is free). Pay-as-you-go: cloud sessions ~$0.06/hour (per-minute billed), proxy data $10/GB (or $0.20/GB egress proxyless/BYO), and LLM tokens at 1.2x provider rates (or BYO key + a 0.2x orchestration fee). Credit-bundled tiers: Starter $100/mo, Business $500/mo ($400 annual), Scaleup ~$2,500/mo (first-party calculator; some sources cite $2,000), plus custom Enterprise.
Billing mechanics
Multi-component usage metering: (1) cloud browser SESSIONS billed per minute (~$0.06/hr PAYG, dropping to $0.03/hr on Business); (2) BANDWIDTH/proxy per GB ($10/GB PAYG to $7.50 Starter to $5 Business; or $0.20/GB egress if proxyless/BYO proxy); (3) LLM TOKENS at 1.2x provider rates if you use their LLM, OR bring your own Anthropic/OpenAI/Google key and pay the provider directly plus a 0.2x orchestration fee; the V2 agent also offers flat per-step pricing. Monthly tiers ($100/$500/~$2,500) bundle credits and unlock lower session/proxy rates + higher concurrency. All plans cap sessions at 4 hours (240 min).
Cost watchouts
Costs stack across THREE meters - session time + bandwidth/proxy + LLM tokens - so hundreds of concurrent sessions running for hours 'add up fast' at $0.06/hr each; the LLM markup (1.2x their model) is on top of session/proxy; there is NO permanent free cloud tier (just $10 trial credits); per-GB proxy is pricey at PAYG ($10/GB); model the expected session-hours x concurrency x token volume before committing
Variable cost rationale
Fully usage-based and multi-axis - spend scales with session minutes, proxy/bandwidth GB, and LLM token volume simultaneously; concurrency and run length are the big levers, and BYO-key (0.2x orchestration) vs their-LLM (1.2x) materially changes the token line
Additional watchouts
Built for developers (API/SDK-first) - non-technical users wanting no-code should look at Airtop/Make; three stacked usage meters make cost hard to forecast and 'add up fast' at scale; no permanent free cloud tier (only $10 credits); proxy at PAYG is $10/GB; stale third-party '$29/mo' listings are inaccurate - there is no such plan
Overage / add-ons
Pay-as-you-go has no cap - you're billed per minute of session + per GB of bandwidth + token usage as you go; the $100/$500 tiers include credits and discounted rates, with usage beyond the bundle billed at the plan's (lower) rates. BYO LLM key shifts model cost to your provider (Browser Use takes a 0.2x orchestration fee instead of the 1.2x markup).
Sales call required
Mixed (some tiers require a call)
Free / trial
Free tier
Lowest paid plan
~$29/mo (ambiguous — verify)
Commercial notes
Started as an OSS project (2024); one of the most popular browser-automation libraries (79,000+ GitHub stars, 1.3M+ monthly downloads); SF-based; SOC 2 certified; claims Fortune 500 users (Airbnb, Amazon, Google, Microsoft); trains custom browser-automation LLMs (understand page structure/forms/navigation better than general models); competes with Browserbase (managed cloud browsers + Stagehand), Bright Data Agent Browser (enterprise-scale), Airtop (no-code), Browserless (self-host)
Key ambiguities
The pricing page is component-based (per-minute sessions, per-GB bandwidth, token markup, per-step V2, hourly personal box) rather than a simple plan list, which is why a single 'monthly price' is misleading; some third-party listings show a stale '$29-$30/mo' plan (and an older 'Developer $50') that does NOT match the current model; exact Scaleup (~$2,500)/Enterprise terms are custom; v2 (per-step) vs v3 (hosted agent) pricing differs
Cancellation / refund
Open-source library is free (self-run); cloud gives $10 signup credits, then pay-as-you-go (no commitment) or credit-bundled monthly tiers (Starter $100 / Business $500, $400 annual / Scaleup ~$2,500); Enterprise custom; BYO key/proxy reduce vendor markup; no permanent free cloud tier
Support SLA / resale
Community/docs/Discord support (OSS + lower tiers); higher tiers + Enterprise add priority support and volume discounts; SOC 2 certified; managed 'BUX' pilot to set up a team workflow for you; stealth browsers + auto-captcha + proxy management; Python/TypeScript SDKs; integrates with major agent frameworks; custom LLMs trained for browser automation (or BYO model)
Missing data
Pricing is component/usage-based, so a single monthly figure is misleading; Scaleup/Enterprise terms are custom; v2 (per-step) and v3 (hosted) rates differ. The open-source library is free; the cloud has $10 signup credits (no permanent free tier) and pay-as-you-go rates (~$0.06/hr sessions + $10/GB proxy + 1.2x token markup, or BYO key at 0.2x), then Starter $100/mo, Business $500/mo ($400 annual), Scaleup ~$2,500/mo, Enterprise custom.
Related vendors
- AgentQL — A self-healing query language for the web: natural language…
- Airtop — Cloud browsers for AI agents that compile plain English workflows…
- Asteroid — Supervised browser and computer use agents for regulated back…
- Astral — Browser based AI marketing agent that operates social platforms like…
- Automat AI — Managed AI agents that operate computers visually to replace legacy…
- Autotab — General AI agent that uses a mouse and keyboard like a person to do…
Alternatives to Browser Use
The closest documented capability profiles to Browser Use among browser and computer-use agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Browserbase12.0 / 14Fuller documented coverage on Security, Identity & Governance and Testing, Debugging & OptimizationBrowser Use vs Browserbase →
- Stagehand10.5 / 14Adds documented Knowledge Grounding & RAG
- Scrapybara9.0 / 14A lighter documented profile than Browser Use
- Emergence AI12.5 / 14Adds documented Knowledge Grounding & RAGBrowser Use vs Emergence AI →
- Airtop10.0 / 14Fuller documented coverage on Memory & State Persistence and Triggers & Channel Coverage
- Island11.0 / 14Adds documented Knowledge Grounding & RAG
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded