---
name: "speaker-diarization-cleanup"
description: "Review speaker labels while preserving uncertain attribution and original timing. Use when the user needs: Get a corrected transcript plus an explicit label-change ledger."
metadata: {"version": "2026-10-17.1", "category": "voice-audio", "publisher": "AIDiveForge"}
---

# Speaker Diarization Cleanup

Get a corrected transcript plus an explicit label-change ledger.

## Requirements

Provide the diarized transcript and reliable identity cues or authorized reference audio.

## Workflow

1. Keep the original transcript and timestamps intact as the reference version.
2. Distinguish evidence-backed speaker changes from guesses based on writing style or topic.
3. Correct a label only when a named introduction, verified reference or clear continuity supports it.
4. Mark ambiguous overlaps and short turns for review instead of assigning a confident identity.
5. Return the revised transcript and every label change with its supporting evidence.

## Worked example

Read `examples/input.txt` and `examples/expected.txt` before adapting the workflow.

## Acceptance checks

Map S1 and S2 using introductions; leave S3 unresolved without sufficient evidence.

## Limits

This is a review workflow, not a bundled voice-identification model. Do not infer protected traits from a voice.

Worked example included. Model-specific performance and vendor compatibility have not been benchmarked.