LocalFree

Browserbash MCP server

PramodDutta/browserbash

Plain-English browser tests for AI agents: real Chrome, deterministic verdicts. Free, no API keys.

3 GitHub stars2 forks362 npm downloads / month
Source code

Install Browserbash in Claude Code

Claude Code
claude mcp add browserbash -- npx -y browserbash-cli
mcp.json (Claude Desktop, Cursor, VS Code, Windsurf)
{
  "mcpServers": {
    "browserbash": {
      "command": "npx",
      "args": [
        "-y",
        "browserbash-cli"
      ]
    }
  }
}

Frequently asked questions

What does the Browserbash MCP server do?

Plain-English browser tests for AI agents: real Chrome, deterministic verdicts. Free, no API keys.

Is the Browserbash MCP server free?

Yes, most likely. Open source and no paid account or API key is mentioned in the listing. Check the provider for current pricing.

Does Browserbash run locally or remotely?

Local. Runs on your machine (stdio). You install it with npx, uvx, Docker or from source.

How do I add Browserbash to Claude Code?

Run: claude mcp add browserbash -- npx -y browserbash-cli

More browser automation MCP servers

View all

five46

3371

sekharsdet/five46

BYOK, fully local AI agent that tests your app/API and writes a real Playwright spec on success.

LocalAccount needed

Anode background Windows desktop

3

skulitom/Anode

Background Windows desktop for AI agents: native GUI automation, screenshots and app testing.

LocalFree

Jidr

3

AbdurahmanAlmehdi/jidr

See and drive Android emulators on a Mac: screenshot, tap, type, UI tree, logs, root.

LocalFree

Anakin

3

Anakin-Inc/anakin-mcp

Web data for AI agents: scrape, crawl, search, deep research, site monitoring, browser automation

LocalAccount needed

Nightglass

083

mizukaizen/nightglass-mcp

Screenshot API for MCP clients. Capture any URL as PNG with mockup mode support.

LocalAccount needed

ZeroDOM

3

DevHusnainAi/zerodom

Deterministic attack-surface perception for AI security agents. Attaches over chrome.debugger to the Chrome you're already logged into (cookies, MFA and SSO already done) and turns the live page into a short numbered interaction graph an agent can act on ([03] button 'Sign In'), with CSS selectors kept server-side and out of the model's context. Cross-identity tools (compare_identities, replay) flag a byte-identical response across two logged-in users as a cross-tenant IDOR, and scan maps the attack surface from the CLI. No LLM in the parse; 71% fewer tokens than Playwright's ARIA snapshot…

LocalFreePython