I’ve been playing around with AI and wanted to share something I built. It’s a server that lets AI language models (like Claude) tap into Trakt.tv data. I’m calling it MCP Trakt.
It’s really just an AI experiment to see how AIs can access real-time info and even your personal viewing history from Trakt.
So, what does it do?
Basically, this server creates a bridge between an AI and Trakt.tv. This means an AI can:
- See what’s trending: Ask it about the hottest shows or movies right now.
- Look up your own stuff: Find out when you last watched a show or how many times you’ve seen a movie on your Trakt account.
- Get details: Grab ratings, read comments, and see what the community’s saying.
- Even check in: You can mark episodes as watched using your AI.
A quick peek at how it works (no need to get too technical):
The MCP part is essentially a way for AIs to “talk” to other systems and get information that wasn’t part of their original training. So, my server lets an AI “talk” to Trakt.tv.
This means an AI can fetch:
- Public Trakt data: Trending shows, popular movies, comments, ratings, and more.
- Your personal Trakt data (if you log in securely): Your watched history, last watched dates, and even let you check in to shows.
A few things you could ask an AI using this:
- “What shows are trending?”
- “Show me my watched movies.”
- “What are people saying about ‘The Godfather’?”
- “Check me in to Season 2 Episode 5 of Breaking Bad.”
If you’re into the technical side or just curious, the project is on GitHub. I’m just sharing this as a fun exploration of what’s possible with AI.