What’s New

browser-use

AI

🌐 Make websites accessible for AI agents. Automate tasks online with ease.

Latest 0.13.7 · by browser-useWebsitebrowser-use/browser-use

Changelog

0.13.7

Changed
  • Stop auto-inheriting output_model_schema into per-page extraction
  • Simplify cross-origin iframe minimum size
Fixed
  • Do not send default params in gemini genconfig
  • Fix tool name in flash prompts from replace_file_str to replace_file
  • Handle file:// URLs in _extract_start_url
  • Fix clear-to-empty on React controlled inputs using native value setter
  • Gate the skeleton-screen 'still loading?' hint on in-flight network requests
  • Fix nested cross-origin iframe extraction filtering
  • Keep selector indices unique across CDP sessions
  • Bound DOM fanout and preserve recoverable state in browser
  • Fix compact cross-origin payment field extraction
  • Stop reusing cached DOM after browser-state timeouts
  • Allow checkbox-sized cross-origin controls safely
  • Call the model with minimal state after capture timeout

Ships the latest version of the CLI, which includes bug fixes and better browser management.

What's Changed

Full Changelog: https://github.com/browser-use/browser-use/compare/0.13.6...0.13.7

0.13.5

Added
  • Accept bu-qa-1 model alias in ChatBrowserUse
  • Support MCP registry
What's Changed

Full Changelog: https://github.com/browser-use/browser-use/compare/0.13.4...0.13.5

0.13.4

Added
  • MCP server for CLI 3.0 with shared MCP runner and max_dim guard
  • Detect LLM output truncation instead of failing with misleading parse errors, with fallback LLM switching on truncation
Changed
  • Pin Browser Harness to 0.1.5 for Windows reliability fixes and cloud-browser suggestion
  • CLI polish including shim for legacy CLI commands, hints for unknown commands, and faster startup
  • Restructure README quickstart around the CLI and Python library paths
Fixed
  • Navigation readiness detection with per-target lifecycle event storage and handling of same-document navigations without timeout burning
  • Markdown extraction no longer destroys URLs and drops long link lines
  • Keep plain BrowserErrors recoverable and preserve structured error memory through execute_action
  • Visibility check no longer mutates shared snapshot bounds
What's Changed
  • MCP server for CLI 3.0: shared MCP runner, max_dim guard.
  • Fix navigation readiness detection: per-target lifecycle event storage, and handle same-document navigations without burning the readiness timeout.
  • Fix markdown extraction destroying URLs and dropping long link lines.
  • Detect LLM output truncation instead of failing with misleading parse errors, and let truncation errors switch to the fallback LLM.
  • Keep plain BrowserErrors recoverable and preserve structured error memory through execute_action.
  • Fix visibility check mutating shared snapshot bounds.
  • Pin Browser Harness to 0.1.5 (Windows reliability fixes, cloud-browser suggestion).
  • CLI polish: shim for legacy CLI commands, hints for unknown commands, faster startup.
  • Restructure README quickstart around the CLI and Python library paths.

Full Changelog: https://github.com/browser-use/browser-use/compare/0.13.3...0.13.4

0.13.3

Added
  • Launch Browser Use CLI 3.0 powered by Browser Harness
  • Add browser-use skill so coding agents can install the Browser Use skill directly
Changed
  • Pin Browser Harness to 0.1.4 and ship the Browser Use skill with the package
  • Improve skill install behavior across Claude Code, Codex, Cursor, Gemini, OpenCode, and related agent skill directories
What's Changed
  • Launch Browser Use CLI 3.0 powered by Browser Harness.
  • Add browser-use skill so coding agents can install the Browser Use skill directly.
  • Pin Browser Harness to 0.1.4 and ship the Browser Use skill with the package.
  • Improve skill install behavior across Claude Code, Codex, Cursor, Gemini, OpenCode, and related agent skill directories.

Full Changelog: https://github.com/browser-use/browser-use/compare/0.13.2...0.13.3

0.13.2

Added
  • Add BU3 browser use models
Changed
  • Accept provider-prefixed models in ChatBrowserUse
  • Default beta example to openai/gpt-5.5
Removed
  • Remove bu-3 and bu-3-max models
What's Changed

Full Changelog: https://github.com/browser-use/browser-use/compare/0.13.1...0.13.2

0.13.1

Added
  • Support Claude Fable 5 model
Changed
  • Anthropic tool choice set to auto when thinking (cannot force it)
Fixed
  • Fix packaged agent tools environment
What's Changed

Full Changelog: https://github.com/browser-use/browser-use/compare/0.13.0...0.13.1

0.13.0

0.13.0 - Rebuilt in Rust [beta]

Added
  • Introduce a new Rust-backed beta agent that provides modern models a more direct browser control loop guided by robust helpers
  • Add x402 skill
  • Add browser-use-core binary discovery
Changed
  • Clean up Rust SDK server implementation
Fixed
  • Fix prerelease version upgrade check
  • Close beta SDK browser resources properly

Browser Use 0.13.0 introduces a new Rust-backed beta agent.

It gives modern models a more direct browser control loop, guided by robust helpers instead of brittle browser abstractions.

uv add "browser-use[core]"
from browser_use.beta import Agent

The existing Python agent remains unchanged.

What's Changed

Full Changelog: https://github.com/browser-use/browser-use/compare/0.12.9...0.13.0

0.12.8

Added
  • Add client header to GoogleChat
  • Add OpenRouter pricing fallback for token costs
Changed
  • Default ChatBrowserUse to bu-2-0
  • Recommend gemini-3-flash-preview in examples and tests
  • Freeze HistoryItem and lock byte-prefix property
  • Move per-step metadata out of agent_state into a tail block
  • Move user request before agent history
Fixed
  • Restrict unix socket file to owner-only access in daemon
  • Refuse evaluate() on restricted browser profiles
What's Changed
New Contributors

Full Changelog: https://github.com/browser-use/browser-use/compare/0.12.7...0.12.8