Build Custom MCP Apps with Strapi To Expose Your Content To Humans and Agents with TanStack AI

You can find all resources here https://strapi.link/strapi-call-tldr Strapi's MCP server now lets plugins register their own tools. In this video, I show how I use that to build custom AI apps on top of Strapi content, for both the people on my team and AI agents. I walk through two plugins I built, TanStack AI (an AI chat inside the Strapi admin panel) and YouTube Transcript. I show how to control which tools each admin role can use, and how to share only some of them with outside AI clients through an admin API token. Then I connect the same tools to a client running a local open source model, and to Grok, where one bot fetches a video transcript and another fact-checks it. Chapters 0:00 What an MCP tool is 1:23 Defining tools in your plugins 2:32 Registering tools with the Strapi MCP server 3:10 Demo: TanStack AI chat in the admin panel 5:21 Tool permissions with admin roles 6:04 Exposing tools through an admin API token 6:57 Connecting a local model client 8:03 Grok bots: fetch a transcript, then fact-check it 10:12 Wrap-up: making Strapi agent-friendly Links TanStack AI plugin: https://github.com/PaulBratslavsky/strapi-plugin-tanstack-ai Strapi docs: https://docs.strapi.io