Plain-English objectives
Describe a browser task in plain English, and BrowserBash turns it into real browser actions without writing selectors or page-object code.
BrowserBash is a natural-language browser automation CLI that turns plain-English goals into real browser actions. Markdown tests, CI output, free local models, no API keys.
BrowserBash is a natural-language browser automation CLI for running browser tasks and tests from plain English. It turns an objective into real browser actions, then returns structured output that can be consumed by CI systems and other automation tools.
The product is built around an open-source CLI, Markdown-based test files, and support for real browsers on local machines or supported cloud grids. The source emphasizes a free default path with local models, no required API keys, and optional dashboard and cloud features for run history and replay.
Describe a browser task in plain English, and BrowserBash turns it into real browser actions without writing selectors or page-object code.
Store scenarios in committable `*_test.md` files, reuse flows with `@import`, and template values with variables for repeatable tests.
Run on free local models through Ollama or free OpenRouter models without API keys, with optional support for your own Anthropic or OpenRouter key.
Drive local Chrome, any CDP endpoint, or cloud grids such as Browserbase, LambdaTest, and BrowserStack from the same objective.
Stream one JSON event per line in NDJSON format and use clean exit codes so agents and CI systems can read the result directly.
Capture video and screenshots with `--record`, then view runs in a local dashboard or a free cloud dashboard with replay and history.
Write end-to-end browser checks as readable Markdown files, then run them in CI with exit codes that indicate pass, fail, error, or timeout.
Automate login, checkout, scraping, and other browser workflows by describing the objective in plain English instead of scripting selectors.
Capture video and screenshots for a run, then replay the session in the dashboard when a step fails or you need a shareable record.
Run browser automation locally with Ollama when you want to avoid API keys and keep the session on your machine.
Point the same test file at a local browser or a cloud grid when your team needs to move between laptop runs and hosted infrastructure.
BrowserBash is designed to automate browser tasks from plain English. The source shows it can run objectives such as logging in, adding items to a cart, scraping text, and running Markdown-based tests, then returning structured output and an exit code.
The source says you can install the CLI with npm, run it with local Chromium or another supported browser provider, and use a free local model through Ollama. A cloud dashboard account is optional for run history and replay.
Yes. The pricing page says the CLI, local dashboard, and a cloud account are free, with no credit card required. It also says the only optional paid item is extended cloud data retention beyond the free retention window, which is coming soon.
BrowserBash supports free local models through Ollama and free OpenRouter models with no API keys required. You can also bring your own Anthropic or OpenRouter key, and the source mentions any OpenAI-compatible server such as vLLM or LM Studio.
The source says BrowserBash is open source under Apache-2.0. It can be used commercially, and the CLI is positioned for CI because it emits NDJSON events and uses exit codes to signal pass, fail, error, or timeout.