Browserbash
Plain-English browser tests for AI agents: real Chrome, deterministic verdicts.
Connector by PramodDutta · GitHub account verified by the MCP Registry
Install
Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.
npx agentmarket install connector/pramoddutta-browserbash
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this server can do and what it asks for, from its tool list and install config. Checked Oct 25, 2026.
- Tools
- 2 read · 6 write · 0 delete · 2 not yet classified
- Secrets
- 1 (API keys, tokens or passwords)
- Install scripts
- Not scanned yet
- Runs as
- Node.js package
Tools by what they do 10
Read from the source code; the running server may list different tools.
Read 2 · Look things up without changing anything.
extractsnapshot
Write 6 · Create, change, send or run things.
clicknavigaterun_objectiverun_suiterun_test_filetype_text
Not yet classified 2 · We could not tell from the name and description.
donewait_for
Secrets and settings it asks for
Secrets: values you paste in that let it act as you.
ANTHROPIC_API_KEY
Files it touches
- Any path the agent passes to its tools
Tools (10)
Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].
clickClick an element. Prefer the ref id from the latest snapshot; a CSS selector or exact visible text also works.
doneFinish the run. Call when the objective is complete (status=passed) or cannot be completed (status=failed). Include a one-paragraph summary.
extractRead the text content of an element and store it under a variable name in final_state. Use when the objective says to extract, read, or "store as".
navigateNavigate the browser to a URL. Call this first when the objective mentions a site.
run_objectiveRun a plain-English browser objective in a real browser and return the structured verdict. Use this to VALIDATE web work: navigate, interact, extract values ("store the heading as 'h1'"). The result JSON has status (passed|failed|error|timeout), summary, final_state (extracted values), and duration_ms.
run_suiteRun a folder of *_test.md files in parallel with memory-aware scheduling. Returns the suite tally (passed/failed/timeout/infra/skipped), per-test verdicts, and estimated cost.
run_test_fileRun one committable *_test.md file (plain-English steps) and return the structured verdict, including deterministic assertion results when the file uses Verify steps.
snapshotGet an accessibility-oriented snapshot of the current page: URL, title, and visible interactive elements (links, buttons, inputs) with stable ref ids. Call this before interacting so you know what is on screen.
type_textType text into an input element, replacing its current value. Optionally press Enter afterwards.
wait_forWait until an element matching the selector or text is visible (up to 15s). Use after navigation or actions that load content.
Signals
- Stars
- 3
- Forks
- 2
- Open issues
- 0
- Last commit
- Jul 10, 2026 (3 months ago)
- Last release
- Jul 9, 2026 (3 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- Registry name
io.github.PramodDutta/browserbash- Version
- 1.5.1
- Auth
- API key
- Packages
- npm browserbash-cli
- License
- License unknown
- Repository
- github.com/PramodDutta/browserbash
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
More from PramodDutta
Cite this page
Browserbash MCP server. Agent Mart. https://agentmart.dev/connectors/pramoddutta-browserbash (last updated 2026-10-24)