From $10 per user/month · Free plan🇺🇸
MCP server and API from Upstash that give AI coding assistants current, version-specific library documentation and code examples.
Playwright MCP overview and key facts (October 2026)
Microsoft MCP server that lets AI agents control browsers through the accessibility tree instead of screenshots.
Free
What is Playwright MCP?
Playwright MCP is Microsoft's Model Context Protocol server for browser automation that lets AI agents and coding assistants navigate and operate web pages using Playwright's structured accessibility tree.
Instead of sending screenshots to a vision model, Playwright MCP gives the model a structured snapshot of the page based on the accessibility tree. The README notes that coding agents working on large code bases may be better served by the Playwright CLI, because MCP loads tool schemas and accessibility trees into the context. It works with clients including VS Code, Cursor, Claude Desktop, Windsurf, Copilot, Goose and LM Studio, and supports Chrome, Firefox, WebKit and Edge. A vision capability can be enabled for coordinate-based clicks. The README states that the allow and block lists for origins are not a security boundary. The GitHub API showed about 37,600 stars and a push on September 25, 2026.
Playwright MCP AI features and use cases
- Tests web apps from a coding assistantA coding agent opens the application, clicks through a flow and checks the result after changing code. Developers verify fixes without writing a separate test script first.
- Automates routine web tasksAgents fill in forms or collect information from internal web tools by reading page structure. Repetitive browser work can be delegated and then reviewed.
- Works without a vision modelBecause the server passes structured data, standard language models can operate pages. Teams avoid the cost and latency of screenshot analysis for most tasks.
What sets Playwright MCP apart
- Uses the accessibility tree rather than pixels, with vision mode only as an option.
- Maintained by Microsoft alongside the Playwright testing framework, with support for four browser engines.
Where Playwright MCP is strongest
United States Available internationally
Official vendor server, usable with MCP clients in any country where the connected service is available.
Questions to ask before you buy Playwright MCP
According to the README, the allow and block lists for origins are not a security boundary, and by default the server uses a persistent browser profile like a regular browser. Use the isolated mode and never give agents access to logged-in sessions they should not use. Page content goes to the model provider of your AI client. Check the terms of websites you automate.
Playwright MCP reviews
No reviews yet. Be the first to share how your team uses it.
Every review shows its verification level. We read each one before it goes live. How we check reviews
Questions about Playwright MCP
Does Playwright MCP need a vision model?
No. It works on structured accessibility data; a vision mode for coordinate-based interaction is optional.
Which browsers are supported?
Chrome, Firefox, WebKit and Microsoft Edge, in headless or visible mode.
Is it free?
Yes, it is released by Microsoft under the Apache 2.0 license.
Playwright MCP update historyno updates yet
No changes recorded for Playwright MCP yet.
Changes the vendor made and confirmed user reports, each checked against the vendor's own pages. We review every tool monthly.
Alternatives to Playwright MCP
Tools in the same category (Developer & cloud tools) with the most overlapping markets.
From $29 per month · Free plan🇺🇸
Managed MCP gateway that connects AI agents to 1,500+ apps with handled OAuth, a free tier and usage-based pricing.
Free🇺🇸
GitHub's official MCP server that lets AI assistants read and act on repositories, issues, pull requests and Actions.
MCP server🇺🇸
Sentry's official MCP server that lets coding assistants query errors, issues and traces and use Seer analysis.