persona-garden-patch

• has_status::[[Seed Stage]]
• in_domain::[[Deep Context Architecture]]
• in_precinct::[[Household Precinct]]

Transcript

A transcript is a cleaned speech-to-text record derived from a meeting recording. Transcripts are always sidecars to their meeting note — they do not exist standalone. The derived_from:: predicate links back to the parent meeting.

- is_a::[\[\[Transcript\]\]](Transcript.html)
- derived_from::[[Meeting Note Name]]↑

Conventions

• Speaker normalization: Full name on first occurrence (**Peter Kaminski**:), first name only thereafter (**Peter**:).
• Cleanup levels: Light (minimal edits), Medium (filler calibration ~10-20% retained for natural rhythm), Heavy (near-zero filler). Medium is the default.
• Frontmatter: created, reviewed, brief_summary, tagline, ai-transcription (source format), ai-processing (cleanup tool and model).
• Naming: YYYY-MM-DD Name - Transcript (Cleaned).md within the compound meeting folder.

Sources

Transcript conventions from the transcript-cleanup and meeting-capture skills. Processing pipeline: raw VTT/SRT → speaker normalization → filler calibration → cleaned markdown.

Relations

• derived_from::[[Meeting Note]] — every transcript belongs to a meeting
• relates_to::[[Sidecar]] — transcripts are companions to meeting notes, similar to how sidecars are companions to binary artifacts

This site is open source. Improve this page.