Review a transcript
import DataBoundary from ‘../../../components/DataBoundary.astro’;
Outcome
Section titled “Outcome”You can assess whether the transcript is good enough to support a summary or question-answering workflow.
Prerequisites
Section titled “Prerequisites”- A session that has completed transcription or is showing live transcription progress.
- Open the session from the session browser.
- Wait for Processing to finish and inspect the transcript panel.
- Compare names, numbers, decisions, and sensitive details against the recording or another authoritative source.
- Add missing title, date, participant, or description context in the session information drawer before generating a final summary.
- If the transcript has no usable speech, retain the source recording and retry only after choosing a more appropriate model or clearer source.
Expected result
Section titled “Expected result”The transcript is available in the session and can support its summary, per-session chat, and indexing workflows.
Data boundary
Section titled “Data boundary”The Whisper transcription path is local. Selecting an External API affects later AI operations, not the local transcription step.
Common problems
Section titled “Common problems”- No speech detected: use a source with clearer speech, less silence, or a different local Whisper model.
- Text is inaccurate: do not rely on a generated summary to correct it automatically; review the original recording and choose the model and language setting deliberately.
- You need speaker labels: Synthezia does not currently provide a verified speaker-diarization or relabeling workflow.