Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Server-side 3D layout: agents can now share research sessions without manually calculating node coordinates. The backend auto-generates semantic layouts via UMAP when the nodes array is omitted.
What This Skill Does
A semantically indexed corpus of 109+ podcast feeds (~7K episodes, ~1.9M paragraphs) exposed as a vector search API. Agents can find timestamped clips from podcasts like Joe Rogan, Huberman Lab, Lex Fridman, and Bloomberg without transcribing audio or stuffing full transcripts into context. Results include playable audio deeplinks and can be assembled into shareable interactive research sessions.
Returns exact timestamped clips with playable audio deeplinks rather than requiring transcript generation, full-context stuffing, or sifting through web search results that return summaries instead of source material.
When to use it
- Finding what Huberman said about a specific health topic across all his episodes
- Researching expert Bitcoin opinions from TFTC and other crypto podcasts
- Locating all podcast appearances of a specific person or company
- Building a shareable research session with timestamped audio clips
- Ingesting a new podcast RSS feed for immediate semantic search
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: What has Andrew Huberman said about cold exposure and dopamine?
- 1Search the corpus semantically for 'Huberman cold exposure dopamine'
- 2Retrieve top matching paragraphs with episode timestamps and audio deeplinks
- 3Create a research session grouping the relevant clips
- 4Return the session URL where the user can play each clip inline
A shareable research session link with timestamped, playable audio clips from Huberman Lab episodes covering cold exposure and dopamine
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.