MCP tools give AI agents capabilities, but prompts tell them how to use those capabilities effectively. In this video, I'll show you how to add MCP prompts to your Zuplo project to create reusable workflows that combine multiple tools together.
Using a bookmark manager as an example, you'll see how to:
- Create a prompt module that returns structured messages
- Expose that prompt as an MCP endpoint
- Test it in MCP Jam and Claude Desktop
- Build a "Research Roundup" prompt that summarizes recent bookmarks and suggests related reading
Prompts let you encode your domain expertise directly into your MCP server, so users get value without needing to know how to prompt the agent themselves.
Try it in Zuplo, for free: https://zuplo.link/check-us-out
Using a bookmark manager as an example, you'll see how to:
- Create a prompt module that returns structured messages
- Expose that prompt as an MCP endpoint
- Test it in MCP Jam and Claude Desktop
- Build a "Research Roundup" prompt that summarizes recent bookmarks and suggests related reading
Prompts let you encode your domain expertise directly into your MCP server, so users get value without needing to know how to prompt the agent themselves.
Try it in Zuplo, for free: https://zuplo.link/check-us-out
- Catégories
- prompts ia
- Mots-clés
- API


Commentaires