---
title: "Transcript Slide Synthesis Loop"
category: "harnesses"
status: "seeded"
sourceLabels: ["Livestream resource", "Slide evidence", "Source map", "Source rules"]
---
# Transcript Slide Synthesis Loop

## Overview
A repeatable way to combine schedule metadata, transcripts, slide OCR, reconstructed slides, and topic pages without promoting weak evidence into confident claims.

## Conference Context
- This wiki has stronger media evidence than Miami: transcripts, stage frames, reconstructed crops, dense slides, and OCR audits.
- The source-boundary page requires official facts, supporting video context, transcript evidence, and OCR evidence to stay labeled.
- Several evaluation talks emphasize that outputs need traceable evidence, not only polished summaries.

## Implementation Pattern
- Anchor the page with official schedule facts first.
- Add transcript-backed claims only when the transcript is linked and speaker/session matching is clear enough.
- Use reconstructed slide crops before full-stage OCR when slide text matters.
- Mark OCR-only claims as OCR-derived unless reviewed against the image.
- Regenerate dependent topic, tool, question, and evaluation indexes after source changes.

## Evidence
- [[source-boundary]] - Source rules
- [[agent-source-index]] - Source map
- [[slide-library]] - Slide evidence
- [[reconstructed-slide-library]] - Slide evidence
- [[dense-slide-library]] - Slide evidence
- [[worldsfair-2026-livestreams]] - Livestream resource

## Evidence Boundary
This is a reusable workflow synthesized from the linked conference evidence. Treat it as a recommended implementation pattern, not as a direct quote from any single talk.