Your First Meeting
Set up audio and models, record a meeting, and review the result.
Last updated: October 24, 2026
Summit records audio on your Mac, turns speech into a transcript, and uses an AI model to create a summary. You can review and edit each stage separately.
The Mac app requires an Apple Silicon Mac and macOS 15 or later. Download it from the download page, then complete onboarding to choose storage, audio access, and models.
1. Choose what to record
Open Settings → Recording and select your microphone and default recording mode:
- Microphone Only for an in-person conversation or a voice note.
- Audio Only for audio played by your Mac.
- Microphone + Audio to capture your voice and the other side of a call.
Allow Microphone access when you use a microphone and System Audio Recording access when you capture Mac audio. Screenshots and screen video have a separate Screen Recording permission; they are optional for audio-only meetings. See recording and importing for permissions and controls.
Tell participants you are recording and follow the rules that apply to the meeting.
2. Set up speech recognition
Open Settings → Speech Models, choose a speech engine, and download a model if you use Whisper. FluidAudio downloads its model on first use. This model turns audio into words; it is separate from the model that writes summaries.
Check the language support for your chosen engine. If you want words to appear during the call, enable live transcription in Settings → Recording or the recording controls. See transcription for engine selection, live transcription, and correction tools.
3. Choose how to generate summaries
Open Settings → AI Models:
- Choose a built-in local model to process text on your Mac after downloading the model.
- Configure a cloud provider if you want to use a remote API. The text needed for the task goes to that provider.
- Connect an OpenAI-compatible local server if you already manage models outside Summit.
In Settings → Processing Flow, decide whether transcription, cleanup, and summary generation should run automatically after recording stops. If you have calls back to back, you can defer processing and run it later.
4. Record the meeting
- Click Record. Recording starts immediately using the selected audio mode.
- Speak normally and check that the audio meters respond.
- Add notes or bookmarks and screenshots as needed.
- Click Stop when the meeting ends.
If automatic processing is on, Summit queues the transcript and summary. Otherwise, use the note’s processing actions when you are ready. You can also import an existing audio or video file.
5. Review before sharing
Open the note and read the transcript alongside the summary. Correct names, numbers, and decisions; AI output can miss or misstate details. Name speakers and check action items before exporting.
Choose a summary template if you want a format suited to an interview, client call, or another meeting type. Regenerate the summary after changing the template or relevant context.
Then put the note in a folder, find related discussions, or export it.
Find the right setting
| What you want to change | Settings page |
|---|---|
| Appearance and startup page | General |
| Calendar access and automatic recording | Calendar |
| Mac and iPhone sync | iCloud Sync |
| Dictation into other apps | Voice Keyboard |
| Microphone, audio mode, live transcription, screen video | Recording |
| Speech engine, language, speech model downloads | Speech Models |
| Local or remote AI for cleanup, summaries, and chat | AI Models |
| Automatic processing, cleanup, speaker identification | Processing Flow |
| Summary instructions | Summary Templates |
| Automatic Markdown export | Export |
| Access from an external AI client | MCP Server |
| Maintenance and repair tools | Labs |
Older releases used AI Provider, Transcription, Audio Recording, and Processing for some of these pages. In the current app, look for AI Models, Speech Models, Recording, and Processing Flow.
If anything gets stuck, use the troubleshooting guide.