Paul Bratslavsky | formerly coding after thirty

I learned how to build MCPs from @kentcdodds, and it's a gift that keeps on giving. I built a local-first knowledge base using @strapijs, @tan_stack, and @ollama with Gemma 4. But if you need to use a frontier model like Claude for more power, just expose your tools and data via MCP.

Getting Started with Strapi MCP 🥳 👉 Watch full video: strapi.link/3QPjhYr @strapijs MCP is now in beta — and it ships by default from v5.47. Here's what the video walks through: 🚀 Enable MCP in http://server.ts and connect Strapi to @claudeai Desktop ⚠️ Admin API + admin tokens run the show — content API permissions don't apply ✅ Register your own services as custom MCP tools Paul even wired two Strapi instances together to chat with his data. 👉 strapi.link/3QPjhYr

How we're approaching Agent Skills with React Router (and Remix) - Put your docs in your `node_modules` - Agent Skill should be small and mostly just pointing to the docs - Full video 👇(I don't have X premium)

Introducing /visual-plan - a skill to generate rich, visual plans for Claude Code and Codex. Plan mode in Claude Code is incredible. But I always find my eyes glazing over when it gives me this huge markdown essay in my terminal. I found I can make much better visual plans with reusable components. So I made a skill called `/visual-plan`. It generates plans as MDX with visual, interactive components. Diagrams, interactive API specs, schema design changes, annotated code, and even pan and zoomable wireframes. So for any UI work, you can look at a wireframe first, comment on it, iterate, and then have the agent work. I’ve found this to be a much more intuitive interface for reasoning about what the agent is doing. It’s somewhat inspired by that popular post about how HTML is better than Markdown. But HTML can be slow and verbose to write. And it doesn’t look good checked into a repo. This has really made me feel like humans and engineering are entering a new abstraction phase, where we reason about things at the plan level. As long as the plan is good, agents are getting more and more reliable at executing on it. Almost to the degree that we trust the C compiler to compile to assembly reliably. Plans are the new intermediate representation. I also made a skill for the reverse of this, called `/visual-recap`. After the agent works, it gives you a recap of everything it did. Same idea: wireframes, interactive API specs and diffs, schemas, annotated code, etc. So now when you’re reviewing what the agent did for you, or looking at a pull request of somebody else’s code, you can see a visual recap instead of just reading a wall of text. It’s all free and open source. You can find it on my GitHub. Will link to it in the reply because we all know how dumb these algorithms are with links.

Services, Connect, Eve, BYOC, and more. Lots of announcements today at Vercel Ship London!

Not sure what hunks are?

Wow, @claudeai can now help me with my video edits. Nice, love how you can expose tools via MCP. This project is local-first and uses a local LLM, but if you want to expose your tools to more powerful models via MCP, you can.

After chatting with @kentcdodds about the power of MCP and his course, I became obsessed with building AI local-first apps and exposing tools via MCP so users can bring their data to use with frontier AI models.

My side project, a local-first video editor built with @TauriApps and @tan_stack on top of local LLM models, to help me create rough-cut videos and export to DaVinci Resolve, now supports using Claude to edit your videos via MCP running in the terminal when you want to use the frontier model.

Claude Code: Auto mode Auto mode lets Claude safely complete long-running tasks without asking permission for every action. But how does it work? 4 key terms: • Classifier: a separate model that approves or denies each action based on your intent and scope • Filtering: the classifier only sees your messages and Claude's tool calls, not Claude’s reasoning or tool outputs, so it isn't influenced when making decisions • Tier check: read-only and easily-undone actions (like edits inside your repo) skip the classifier; everything else goes through it • Probe: a server-side check that scans incoming content, like web pages, for prompt injection and flags it to Claude I'm working on a longer video that explains how auto mode works and how to configure it. What questions do you have about it?

We've been working on TanStack AI for a while now and I wanted to help onboard people into this wonderful world of AI powered apps! 🚀 I'm going to be holding a workshop on building AI powered apps and I would love for you to join me if you want to learn how you can add AI into your products! You can find the workshop details on the link below: epicweb.dev/events/buildin…

@AlemTuzlak is awesome.

The Strapi MCP opens endless possibilities.

@tan_stack desktop app using @TauriApps as a wrapper, connected to @Ableton live via Max for live based on Max by @cycling74. Used Claude to help me build the POC.

I'll be doing a workshop at @thejsnation on building AI powered apps, and while I was working on the workshop I said to myself: "Why not make this an Epic Workshop" And that's exactly what I'm going to do! If you're interested I'll be announcing the workshop soon!

Strapi 5.47.0 is out. This is what this crazy community is shipping 8 days after 5.46.1 🤯🤯🤯

HERE WE ARE: The Strapi MCP beta is live 💥 Can't wait to see what you will 𝚋̶𝚞̶𝚒̶𝚕̶𝚍̶prompt with it.