Post

You can store mcp config inside `.mcp.json` in the project root as per docs.anthropic.com/en/docs/clau.... Claude will ask you to approve it first time you run it. `.mcp.json` can be checked into source control and shared with the rest of the team

easily one of my favorite mcp servers! I have even started creating end to end tests as markdown files and claude code will add notes on the results as it runs the tests.

Excellent! Thank you

The MCP is so helpful! One thing I've found is it eats up a ton of context on products with large-ish DOMs. So if you're planning on running a flow several times (say for testing), ask the MCP to write an actual Playwright .py that you can run over and over.

Decentralized tools can leverage browser automation, enhancing Claude's capabilities in tasks like web scraping and automated testing.

Useful pattern. For business checks agents can also discover without hand-wiring: jithox.com/connect-your... + jithox.com/llms.txt — free no-account tools on mcp.jithox.com/mcp

The context cost is the part that bites after the demo: snapshot-heavy pages make repeated runs expensive and harder to diff. I usually generate a small deterministic Playwright harness for the stable path, then keep MCP for exploration and the odd weird edge case.

Oh interesting, maybe there's a Playwright install command I had run in the past

You can use it also with cursor, and there is something really unnerving in seeing the model navigating on it's own. (it's better at following instructions that most people but it will get confused and fail to some some tasks anyway...)

I also did this today! It worked like 30% of the time for me though

Took me so long to figure this out - I was wrestling with configs and trying to install globally before I eventually realised it's literally just this command and then it works 😅

📌