Tabnine doesn't expose MCP — use the npm install path and rely on Tabnine's codebase indexing.
Tabnine's flagship product is its autocomplete and Chat — it doesn't currently expose a public MCP-server or custom-skill registration. Your best play with Tabnine is to install CleenUI normally and let Tabnine's codebase awareness pick up the patterns from your imports. This page covers that flow.
Without MCP-style skill registration, install CleenUI like any other library. Tabnine's chat-with-codebase will pick up your imports.
Add the tokens.css and styles.css imports manually — this is the part the Setup skill normally automates, but you can do it once by hand.
Now that CleenUI is in your project, Tabnine Chat can pattern-match from existing imports. Ask it to build new screens using the components.
Open Tabnine Chat and ask 'what's in package.json'. If @cleenui/react is in the answer, Tabnine has the dependency in context.
Tabnine's suggestions improve as you use the library more. Add one or two CleenUI imports manually, then Tabnine will pattern-match from those.
Two paths to your first component. Pick the one that fits how your team builds.
One prompt to your AI tool. The Setup skill handles dependencies, design tokens, build config, and component registration — all without leaving your editor.
The classic flow. Install the package, import the styles, drop in your first component. No agents required — same end result.