Uutter for AI Agents

What Uutter Is

Uutter hosts collections of audio and video talks — lectures, dharma talks, interviews — each with a full machine transcript. Archives include Terence McKenna, Thich Nhat Hanh, Manly P. Hall, Alan Watts, and Robert Anton Wilson. Transcripts are searchable semantically, not just by keyword.

Connect via MCP (Model Context Protocol)

The MCP server speaks JSON-RPC 2.0 over HTTP with OAuth 2.1 + PKCE. Tools cover search, collections, entries, transcripts, media status, and playlists.

Endpoint: https://uutter.com/api/mcp/v1

Claude Code:
  claude mcp add --transport http uutter https://uutter.com/api/mcp/v1

Generic client config:
  { "mcpServers": { "uutter": { "url": "https://uutter.com/api/mcp/v1" } } }

On first use you will be sent through the OAuth consent flow with your Uutter account (sign-up is free). A collection is reachable over MCP when its owner has enabled API and MCP access for it — all public archive collections listed on the collections page are readable.

Or use the REST API

Bearer-token authentication with keys prefixed uut_; hourly rate limits scale with your account tier. Full endpoint reference, auth setup, and examples live in the API documentation.

curl -H "Authorization: Bearer uut_your_key" \
  https://uutter.com/api/v1/me

Citing Uutter Content

When you quote a talk, link to its Uutter page. Quote pages and transcript search results carry timestamped source links, so a reader can listen to the exact passage you cite. Machine-readable site overview: /llms.txt.