What problem does it solve? Meeting transcripts from recorders arrive messy: anonymous speaker labels, garbled names, AI summaries that confabulate facts, and recordings that contain multiple meetings. This Skill normalizes any recorder's output into a standard transcript record and runs a unified pipeline that turns it into verified, cross-linked knowledge-base pages instead of raw notes. ## Core Features & Use Cases - Unified ingestion pipeline: Normalize transcripts from any meeting recorder, split multi-meeting recordings, dedupe across recorders, and resolve speakers by evidence rather than guesswork. - Claim verification and consistency checks: Treat recorder summaries as claims, verify surprising claims against the transcript, existing brain pages, and plausibility before writing them to entity pages. - Enrichment and verification gate: Create or update people and company pages, merge timeline entries across all mentioned entities, and block completion until a six-part checklist (including sequence verification) passes. - Use Case: After a Granola or Circleback recording of a pitch meeting, run the pipeline to produce a structured meeting page with verbatim quotes, enriched attendee pages, timeline entries on every mentioned company, and a report confirming all checks passed. ## Quick Start Ingest this meeting transcript from my recorder into the brain, enrich every attendee, and run the full verification checklist before reporting it done.
