USER GUIDE / 07
Proofreading & summaries
Retain the recognition transcript and treat AI output as suggestions to review.
Screenshots show the app interface with example settings and sessions. Use your own models, services and recordings.
First time here? Start with a taskFollow illustrated tutorials from installation and model setup to your first result →Configure and run
- Configure an available LLM service, endpoint, model and authentication in API or advanced settings. A translation model may not be appropriate for long summaries.
- Use the appropriate proofreading or summary action in Raw Text or the meeting area, wait for completion and inspect any error.
- Compare names, numbers, negation and quotations with the transcript before saving or exporting.
- 1Copy the source transcript.
- 2Export the transcript as a text file.
Three different results
| Feature / message | Guidance |
|---|---|
| Proofreading | Correct wording and recognition mistakes; compare each passage with its source. |
| Summary | Condense the key information; it does not retain every original sentence. |
| Minutes / LingLu reports | Organize sections, decisions and actions; use their dedicated generation, saving and export workflows. |
- 1Check the summary against the transcript.
- 2Check actions, owners and dates.
Retain existing content on failure
For timeouts, authentication errors or empty results, keep the transcript and existing drafts and retry after correcting configuration. Verify the selected session. After switching sessions, do not overwrite the new session with a previous draft.