BrowserBash

BrowserBash is a free, open-source command-line tool that lets you automate a real web browser by writing a single sentence in plain English. Instead of writing complex test scripts with selectors and locators, you just describe what you want to happen — like "Fill the form and submit" or "Open Hacker News and store the top story title" — and an AI agent carries it out in a real Chrome browser. The CLI is installed via npm and works immediately with no signup, no API keys, and no credit card. A free account adds a dashboard for run history, video recordings, and per-run replays.

BrowserBash is built on a modular architecture with three swappable layers: the browser provider (local Chrome, BrowserStack, LambdaTest, etc.), the engine that interprets your English (Stagehand by default, or a built-in Anthropic tool-use loop), and the LLM that does the thinking. The default stack uses Ollama with a local model — completely free, private, and unmetered. You can also bring your own Anthropic, OpenAI, or Google key, or use any OpenAI-compatible server. The tool outputs NDJSON events on stdout with exit codes (0 for pass, 1 for fail) that your CI pipeline understands natively. It supports markdown test files, secret masking, and a --agent mode for streaming events.

This tool is built for developers and QA engineers who want to write end-to-end browser tests without the overhead of traditional frameworks like Playwright or Selenium. It's also useful for anyone who needs to automate repetitive web tasks — like scraping prices, logging into sites, or checking page content — without writing code. Because it's open source and runs locally, it's a good fit for teams that want full control over their test infrastructure and data privacy. The case study on the site shows how a production Playwright suite was rewritten as a single plain-English file and run on a free local model.