Transcription & Summaries
How ScribeScoop converts your recordings into text transcripts and structured summaries using AI.
Overview
After you record a meeting or conversation, ScribeScoop can process the audio file to produce two outputs:
- Transcript -- A complete text version of everything said in the recording.
- Summary -- A structured summary of the recording, formatted according to your selected summary profile.
Transcription is powered by OpenAI Whisper, a speech recognition model that supports a wide range of languages and accents. Summarization uses a large language model to analyze the transcript and extract the most important information.
Processing a recording
- Navigate to the Recordings tab in the sidebar.
- Select the recording you want to process from the file list. Files that have already been processed are marked with an "OK" badge.
- Choose a summary profile from the dropdown at the bottom of the panel. The default profile is Meeting Notes. See Summary Profiles for details on available profiles and how to create custom ones.
- Click the Process button. Processing begins immediately.
Processing stages
The application shows status updates as it moves through each stage:
| Stage | Status shown | What happens |
|---|---|---|
| Upload and transcription | Transcribing... | The audio file is uploaded to the ScribeScoop processing service and converted to text by the Whisper model. |
| Summarization | Summarizing... | The transcript text is analyzed and a structured summary is generated using the selected summary profile. |
| Complete | Results displayed | Both the transcript and summary are available in the Results tab. |
Viewing results
When processing completes, the Results tab opens automatically and displays two sub-tabs:
- Summary -- Shows the structured summary, rendered with formatting such as headings, bullet points, and numbered lists.
- Transcript -- Shows the full verbatim transcript as plain text.
See the Results Viewer page for details on searching, copying, downloading, and printing your results.
Cached results
After a recording has been processed, the results are cached locally on your computer. This means:
- Viewing a previously processed recording loads the results instantly without contacting the server.
- You can access your results even when offline (as long as they were previously processed).
- Cached results are stored in a local database and persist across application restarts.
Recordings that have cached results show an "OK" badge in the Recordings file list. Clicking on them opens the results directly instead of showing the Process button.
Re-processing a recording
If you want to generate a new summary using a different profile, you can re-process the recording. Select the recording, choose a different summary profile, and click Process again. The new results will replace the cached version.
Processing limits
Each subscription plan includes a monthly allowance of transcription minutes. Processing a recording uses minutes from your allowance based on the duration of the audio file. Summarization does not consume additional minutes.
You can check your remaining minutes in the sidebar usage bar or on the web dashboard. See Billing & Plans for details on plan limits.
File size limit
Audio files up to 100 MB in size can be uploaded for processing. For very long recordings, using MP3 format is recommended to keep file sizes within this limit.
Error handling
If processing fails, an error message is displayed in red below the process button. Common causes include:
- Network issues -- Check your internet connection and try again.
- Authentication errors -- Your session may have expired. Sign out and sign back in.
- Usage limit reached -- You have used all available transcription minutes for the current billing period.
- File too large -- The recording exceeds the 100 MB upload limit.