What does BrowserAct cost?

Almost everything is free with no signup required. Only two features require payment: managed proxies (Dynamic/Static) and stealth browsers beyond the first 5.

Which AI agents can use BrowserAct?

BrowserAct works with any agent that can execute shell commands and load Skills, including Claude Code, Cursor, VS Code, OpenCode, OpenClaw, Codex, and Gemini CLI.

What is Skill Forge and how does it work?

Skill Forge is a companion tool that explores a website once, discovers its APIs and data patterns, generates a deploy-ready Skill package, and then runs reliably without re-exploration—handling 500 or 5,000 records through the same stable path without manual scraper writing.

Large Language ModelsHacker NewsJul 28, 2026Jul 28, 2026, 19:00 JST

BrowserAct: Open-Source Browser Automation for AI Agents

Send on LINE

3 Key Points

What happened
BrowserAct, a browser automation CLI designed for AI agents, was released as open-source software. It enables agents to bypass anti-bot protections, hand off control to humans when stuck, run parallel tasks without interference, and isolate multiple accounts in independent browsers.
Why it matters
AI agents often struggle with websites that block automated access or require human intervention. BrowserAct addresses this by offering three layers of anti-blocking (fingerprint spoofing, CAPTCHA solving, and human handoff), three browser modes for different scenarios (Chrome login reuse, stealth scraping, and logged-in multi-browser parallel work), and a token-efficient interface designed for LLM reasoning rather than human scripts.
What to watch
The tool is free for basic browser automation, stealth extraction, and the first 5 stealth browsers; paid features include managed proxies and additional stealth browsers beyond that. It works with Claude Code, Cursor, VS Code, and other agents that can execute shell commands. A companion tool, Skill Forge, auto-generates reusable data extraction Skills for repeated scraping tasks across the same websites.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

BrowserAct addresses a practical friction point in AI agent deployment: most websites employ anti-bot defenses that block standard automation tools, and agents often lack a graceful way to escalate to human operators when they encounter obstacles. The tool's architecture reflects this reality by layering three distinct blocking-defeat mechanisms—environment (fingerprint spoofing and proxy rotation), execution (CAPTCHA solving and protected-page extraction), and human (live handoff URLs)—rather than betting on a single approach.

The design philosophy shifts away from human-oriented browser scripting and toward agent reasoning. Rather than outputting verbose JSON or HTML that consumes tokens and requires DOM parsing, BrowserAct returns compact indexed text ("click 3 / input 2") and attaches semantic memory to each browser session for task matching. This reflects the constraint that LLMs work better with concise, structured state than with raw page markup.

The companion tool, Skill Forge, extends the utility by automating the most tedious part of integration: discovering how each target website structures its data and APIs, then codifying that knowledge into a reusable Skill. This lowers the barrier for agents to scrape at scale without hand-written scrapers.

FAQ

What does BrowserAct cost?: Almost everything is free with no signup required. Only two features require payment: managed proxies (Dynamic/Static) and stealth browsers beyond the first 5.
Which AI agents can use BrowserAct?: BrowserAct works with any agent that can execute shell commands and load Skills, including Claude Code, Cursor, VS Code, OpenCode, OpenClaw, Codex, and Gemini CLI.
What is Skill Forge and how does it work?: Skill Forge is a companion tool that explores a website once, discovers its APIs and data patterns, generates a deploy-ready Skill package, and then runs reliably without re-exploration—handling 500 or 5,000 records through the same stable path without manual scraper writing.

Hacker NewsRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.