About BrowserBash

BrowserBash turns plain-English objectives into real browser actions, with no selectors or code to maintain. It runs on free local models and supports real browsers, including cloud grids like Browserbase, LambdaTest, and BrowserStack. The tool is free and open-source under Apache-2.0 and is designed as a validation layer for AI agents in testing workflows.

The project is built to be local-first and private, operating without API keys by default. It offers a dashboard optional cloud account for run history, recordings, and per-run replay, all accessible at no required cost. The development focuses on accessibility for testers, QA engineers, and teams adopting CI workflows.

BrowserBash provides a structured, machine-readable output format (NDJSON and exit codes) to integrate with CI and AI agents. It supports Markdown test files, secret masking, saved logins, and per-run verifications that map to real browser actions.

The system emphasizes open-source principles, allowing inspection and contribution, and keeps the CLI free to start with optional paid enhancements for cloud retention and additional features. It supports local and cloud execution environments and aims to reduce reliance on API keys or per-token billing.

Use cases

Integrate BrowserBash into CI pipelines with NDJSON output and exit codes.

Key features

Plain-English objectives that drive real browser actions without code or selectors.
Markdown test files with composition and imports
Variables and secret masking
Open-source, zero API keys, free local models
Real browsers and multiple providers
MCP server integration for agent-based testing
Deterministic assertions and CI-ready exit codes

Pros and cons

Pros

Free and open source under Apache-2.0 license.

No API keys required by default.

Cons

Cloud run history retention is optional and limited to 15 days in the free tier.

Pricing

Free

$0 / forever

The full open-source CLI — every command, no locked features
Plain-English automation: an objective in, real browser actions out
All browser providers: local Chrome, any CDP endpoint, Browserbase, LambdaTest, BrowserStack
Free local models via Ollama or free OpenRouter models — zero API keys
Markdown *_test.md tests with @import composition and variable templating
NDJSON agent mode and CI exit codes for pipelines
Session recording with --record (video & screenshots)
Free local web dashboard via the dashboard command
Free cloud account: 15-day run history, recordings & per-run replay
Open source under Apache-2.0 — audit it, fork it

Reviews

Loading reviews…

Frequently asked questions

What is BrowserBash and what does it do?

BrowserBash is a free, open-source CLI that turns plain-English objectives into real browser actions with no selectors or code, running locally or on cloud grids.

Do I need API keys or a credit card?

No. BrowserBash runs on free local models through Ollama or on free OpenRouter models with zero API keys and no credit card.

What browsers and providers are supported?

It drives a real local Chrome and can connect to cloud grids like Browserbase, LambdaTest, and BrowserStack.

Can I use BrowserBash in CI?

Yes. It emits structured NDJSON and standard CI exit codes to gate pipelines.

Alternatives to BrowserBash

GPTConsole

GPTConsole provides autonomous AI agents for developers to build, test, and deploy production-ready apps using agents like Pixie and Chip.

Userdoc

Userdoc turns code and ideas into software specs, user stories, and acceptance criteria to help teams document and plan software with AI.

DiscuroAI

DiscuroAI defines and executes AI workflows for developers, enabling building, testing and consuming complex AI flows via a single API call.

GooseAI

GooseAI provides a fully managed NLP API service intended to reduce AI infrastructure costs for developers.

GitHub Copilot

GitHub Copilot is an AI coding agent integrated with GitHub to help write, understand, and ship software across development tasks.

LangChain

LangChain provides an open agent platform for building, deploying, and governing autonomous agents and observations, targeting teams that build AI-powered workflows.