MCP

AI-agent format

Map a whole field from a conference playlist

Import the conference, generate mindmaps per talk, and see the field's themes and factions in one view.

ResearcherLearning & Upskillinglibrary_import_youtube_channelrun_youtube_tool
The use case

The problem

A field's annual conference is its self-portrait: what the community considers solved, contested and frontier, painted in forty talks. Researchers entering a field — PhD students, industry scientists switching domains, analysts covering a technical space — are told to "watch the recent conference talks," which is genuinely good advice that nobody can follow at forty hours per conference per year.

So field-mapping happens by sampling: watch five talks, extrapolate the rest, miss the theme that only becomes visible when a dozen talks orbit it. The structure of a field — its clusters, its factions, the questions everyone circles — is an aggregate property. You can't see it from samples.

The mechanics

How veridive MCP helps

library_import_youtube_channel ingests the conference — the sessions playlist, one import, transcripts and all. run_youtube_tool's mindmap then renders each talk's conceptual structure: the claims, methods and dependencies, as a map instead of an hour of video.

Forty mindmaps side by side is where the field's shape emerges: which concepts appear in every third talk (the field's core), which cluster together (the factions and schools), what everyone cites as open (the frontier — which is also your research-gap list). summarizer fills in the narrative per talk where a map needs prose. The aggregate view that sampling can't produce, from the full corpus, in an afternoon of reading instead of a season of watching.

Step by step

How to run it

  1. 1
    Connect veridive MCP

    One-time setup: add the veridive MCP server to Claude, ChatGPT, Cursor or any MCP client and generate your API key. Takes about two minutes.

  2. 2
    Import the conference

    The sessions playlist, tagged by event and year — the corpus the map gets drawn from.

    Import this conference's sessions playlist into my veridive library, tagged '<conference year>': <playlist URL> — use veridive's library_import_youtube_channel.
  3. 3
    Render the talks as maps

    A mindmap per session: concepts, claims, methods — the hour, as structure.

    Generate a mindmap for each imported '<conference year>' session — core concepts, main claims, methods used.
  4. 4
    Read the field off the aggregate

    Across all maps: the recurring core, the clusters, and what everyone flags as unsolved.

    Across all these session maps: which concepts recur most, what thematic clusters form, and what do speakers repeatedly name as open problems?
Before vs after

Problems it solves

  • Forty hours per conference makes 'watch the talks' advice unfollowable.
  • Sampling five talks misses the themes that only emerge across dozens.
  • A field's factions and frontier are aggregate properties — invisible per-talk.
  • New-entrant researchers spend a semester on orientation a corpus view delivers in days.
The payoff

Why it's valuable

  • The field's self-portrait, read whole: core, clusters, factions, frontier.
  • The open-problems aggregate doubles as a research-gap list with provenance.
  • Year-over-year re-runs show the field moving — themes rising, questions closing.
  • The imported corpus stays queryable: any later question runs against the same conference.
Copy & paste

Example prompts

Import the '<conference year>' sessions playlist, tagged: <URL> — use veridive's library_import_youtube_channel.
Mindmap per imported session: concepts, claims, methods.
Across the maps: recurring core, clusters, and named open problems?
Questions

FAQ

Which tools do the mapping?

library_import_youtube_channel ingests the playlist; run_youtube_tool's mindmap renders each talk's structure, with summarizer for narrative detail where needed. The cross-talk synthesis is your assistant reading the maps together.

Does this replace actually watching the key talks?

It tells you which three talks are worth your full attention — the map is for orientation, the watching is for depth. Most researchers end up watching fewer talks more deliberately.

How does this compare with the expo-recon use case?

Same import machinery, different reading: expo recon extracts announcements and business signals for a briefing; this reads conceptual structure for field orientation. Researchers want the map; PMs want the memo.

Keep going

Related use cases

Run this in your own assistant

Connect the veridive MCP server to Claude, ChatGPT or Cursor and paste the prompts above — setup takes about two minutes.