Vew Mcp
Drive real Chromium through Playwright and return compact, ref-addressable page snapshots for agent-led UI checks.
Overview
vew-mcp is a MCP server for the Ship phase that drives Chromium via Playwright and supplies compact ref-addressable browser snapshots to AI assistants.
What is this MCP server?
- Browser MCP backed by real Chromium via Playwright—not mocked DOM stubs
- Compact ref-addressable page snapshots sized for LLM context budgets
- stdio npm package @yawlabs/vew-mcp (v0.2.0)
- Lets agents click, navigate, and inspect live pages during ship checkpoints
- Pairs with solo-builder pre-launch sweeps of signup, billing, and dashboard flows
- npm package @yawlabs/vew-mcp at server version 0.2.0
- Transport: stdio; engine described as Chromium via Playwright in server metadata
What problem does it solve?
Agents can patch frontend code but still cannot easily see or step through the live UI, so solo builders re-run manual browser tests outside the chat.
Who is it for?
Indie web builders who want Playwright-grade realism for agent-led QA, repro, and pre-launch click-through checks.
Skip if: Headless API-only products with no web UI, or teams that require hardened multi-tenant browser farms instead of local Chromium.
What do I get? / Deliverables
After registering vew-mcp, your agent can navigate real pages, capture structured snapshots, and support hands-on test flows in one session.
- Live browser interactions and compact ref-addressable page snapshots
- Agent-documented repro steps for UI defects
- Faster pre-ship validation of critical user flows
Recommended MCP Servers
Journey fit
Real-browser verification belongs in Ship when you validate flows before release, not when you are still scoping the idea. Playwright-driven browsing supports manual-ish test passes, repro steps, and UI regression hunting under the testing subphase.
How it compares
Playwright browser MCP with snapshot refs, not a static screenshot skill or marketing SEO crawler.
Common Questions / FAQ
Who is vew-mcp for?
Solo builders shipping web apps who want their coding agent to drive a real browser for testing and debugging via MCP.
When should I use vew-mcp?
Use it in Ship during testing—before release or when reproducing UI bugs—on staging or local URLs you control.
How do I add vew-mcp to my agent?
Install @yawlabs/vew-mcp, register the stdio MCP server in your agent host, and ensure Playwright/Chromium dependencies can run on your machine.