This is an advanced workflow requiring configuration file edits and multiple steps. If you need basic meeting summaries, consider using the standard Meeting Mode. Or, if you want speaker-separated transcripts with basic AI processing, simply enable speaker separation in a Custom Mode with Nova (Cloud) for voice.

​
Overview

This guide shows you how to combine speaker-separated transcription with AI analysis to understand each participant’s contributions in your meetings. You’ll learn how to capture detailed transcripts that track who said what, then use AI to extract meaningful insights from individual speakers’ input.

​
Transcription Setup and Recording

1

Create a Speaker Separation Mode

Create a new Custom Mode with these settings:

  • Enable “Identify Speakers” toggle
  • Enable “Record from System Audio” if recording live meetings
  • Select your preferred voice model
While all voice models support speaker separation, Nova (Cloud) delivers the most accurate and reliable results for this feature.
2

Disable AI Processing

Disable AI processing initially to speed up transcription. This approach is suggested since Superwhisper currently can’t process speaker-separated transcripts with AI in the same mode. Remember to back up your mode’s config file before making changes:

  • Navigate to: ~/Documents/superwhisper/modes/
  • Open your mode’s JSON file in your preferred text editor
  • Change “languageModelEnabled” from true to false
  • Save and restart Superwhisper for changes to take effect
3

Capture & Transcribe Your Meeting

Choose your preferred recording method:

4

Access Your Transcript

  1. Open Superwhisper History
  2. Select the Segments tab
  3. Rename speakers (optional):
    • Click speaker names directly, or
    • Use the right sidebar’s Speakers section
  4. Click “Copy to Clipboard” to get the full speaker-separated transcript

​
AI Processing and Analysis

​
Option A: External AI Services

  • Use OpenAI, Claude, or your preferred AI provider
  • Request your desired analysis on the transcript

​
Option B: Superwhisper Custom Mode

1

Create a new Custom Mode

  • Write an assistant-focused prompt
  • Enable Application Context
    For tips and ideas on how to get the best results with your custom instructions, check these prompting tips.
2

Process Your Transcript

  • Open any text editor
  • Paste your speaker-separated transcript
  • Activate the mode you have just created
  • Dictate your analysis request

​
Sample Procesing Requests

"Summarize this meeting's key points by speaker"
"Extract action items and assign them to each speaker"
"Show which team members contributed to each key decision"
"Create a participation report showing who discussed which topics in this meeting"
"Break down topic ownership and insights by participant from this meeting's transcript"

Was this page helpful?