Castify Releases MCP

Castify just got a lot smarter. Our new MCP connector brings your video library directly into Claude, you can search every video in your Screencastify and Castify account without leaving your workflow. And it's just the beginning: we'll be rolling out new capabilities on an ongoing basis as we build out what's possible when your videos meet AI.

Use cases:

• Match videos to any document
  Drop in a syllabus, lesson plan, or doc and Claude finds the most relevant videos in your library and pulls in the watch links automatically.
• Build watch playlists
  Ask Claude to sequence your video lessons into a logical viewing order, ready to share with students or teammates.
• Organize your library
  Get suggestions for how to structure your folders and categories based on the content you've actually got.
• Power customer support
  Point Claude at a customer question and get back the exact video that answers it  instant, accurate reference without digging through your library.

How to set up your Claude Custom Connector:

1. Go to your Castify Settings page and scroll down to MCP Server Credentials.
2. Create a new client and copy the credentials.
3. Go to Claude’s Connectors page and add a custom connector.
4. Enter Screencastify’s server URL (https://api.screencastify.com/mcp/message) and paste in your credentials.

That's it your Castify connector is ready to use.

What does MCP enable?

The MCP Integration with Claude connects your video content library directly with Claude's AI capabilities. This means you can automatically analyze, organize, and share your videos using smart features, without manual sorting or guesswork. The integration pulls in data such as view counts, quiz completions, comments, and video topics, allowing you to make informed decisions about your content. It also enables you to automate repetitive tasks like organizing videos by subject or creating shareable playlists with detailed descriptions and links.

Coming soon:

We're mapping Screencastify's functionality to MCP so that virtually the entire application can be driven by agents, or plugged into whatever workflow you're already using across all LLMs.

‍