Read alongside the recording
Select a saved meeting from the sidebar. Use its playback controls to listen while reviewing the transcript. Paragraph and word navigation help you locate a passage. A collection has multiple audio timelines; use its source meeting when you need a specific recording position.
Correct the transcript
Enable transcript editing and correct the text. Finish the edit before switching tasks, then check that changes were saved. When a save error appears, keep the meeting open and resolve it before closing the app. New analysis should use the corrected transcript; an older report describes the input it originally received.
Separate and review speakers
- Open Meeting Actions → Speaker Diarization for a saved meeting.
- Prepare the local speaker model if required and start processing.
- Review the proposed segments and speaker assignments.
- Apply the result when it matches the recording. Use the available undo/review workflow if assignments need correction.
Overlapping speech, poor recordings and short samples can affect results. Speaker labels are suggestions to review, not identity proof.
Keep a reusable speaker library
Open Speaker Library to manage profiles. Add a clear single-speaker sample, or use an eligible excerpt from a meeting. Review names and matches before applying them. Speaker samples and voice features stay in the local workspace for this feature.
Keep the source traceable
Source paragraph numbers can be shown while working with reports. These source identities support navigation between conclusions and the transcript, including eligible links in combined PDF exports. Correcting a transcript can change the context of an older conclusion; regenerate or review its report when needed.