Thrive MCP: Full Release

The Thrive MCP server exposes Thrive's platform capabilities to AI assistants through the Model Context Protocol, the same underlying connection that lets external AI tools work with Kiki. Previously available in beta to a small group of early adopters, it is now hardened and generally available, working across the major AI tools customers already use.

• More than 30 capabilities spanning skills and tags, learning content, wiki pages, posts, badges, assignments and audiences
• Read and write access to learning content, pages, posts and tags, and read access to audiences and users
• Works across Copilot, Gemini and Claude

• Removes the need to navigate the platform UI or build against the API for AI-first workflows
• Bulk tag assignment, content drafting and skills updates can be driven by a prompt
• Customers are not locked to one assistant, since it works across multiple tools

• A customer connects the Thrive MCP server to their chosen AI tool, authenticating as themselves
• The assistant interprets a natural language request, calls the relevant capability, and returns the result conversationally
• Actions are scoped to that individual's permissions, not a shared service account

• An admin asks their AI tool to tag an article with the right skills and topics, and the assistant applies them automatically.
• An L&D team asks which badges a user has earned and what is still assigned to them, pulling the answer in one go.

• Admins and content managers use natural language to manage tags, skills and content at speed
• L&D and skills teams query and update user skills data without navigating multiple screens