Performance and your hardware
Daisy runs entirely on your machine. Recording, transcription, and speaker labels never leave your computer. That's the point of Daisy — and it also means your hardware sets the pace.
The expensive part happens while the meeting runs.
Transcription happens during the meeting
Daisy transcribes as you talk, and that text is your transcript. When you press ✦ Finish & summarize, it promotes what it already has. The transcription cost is spread thinly across the length of the call.
The default engine is chosen to make that cheap:
- Windows and Linux — a bundled Parakeet model that runs on your CPU, so it needs no graphics card and leaves the one you have to your meeting app.
- macOS 26+ — Apple's own on-device speech engine, which uses the Mac's dedicated hardware for it.
- Whisper — used when you choose it, or as the fallback. This is the one that wants a GPU: Metal on Mac, Vulkan on Windows and Linux, CPU otherwise. The CPU path works but is markedly slower.
What still takes time when you stop
Four things:
| Step | What it costs |
|---|---|
| Echo cancellation and audio cleanup | Seconds to a minute or two, scaling with meeting length |
| Speaker labels (diarization) | On-device, and quick |
| Filling coverage gaps | Only runs if the live pass missed a stretch — usually nothing |
| The AI summary | Almost always the longest step on a modest machine |
That last row is the one to pay attention to — see below.
When a full transcription pass does run
There are two cases where Daisy transcribes a whole file from scratch, and both are visible as a transcribing… n/m chunks progress bar:
- Imported audio — a voice memo or an existing recording has no live pass behind it. See Importing audio and notes.
- A session recorded with no model installed, or one where the live transcriber was demoted mid-meeting because the machine couldn't keep up. The audio is intact; Daisy just has to do the work now.
This is the case where a GPU and a smaller model both help.
The summary is a separate step
Summaries, chapters, and analysis need an AI model, and on a modest machine that's the slowest part of the whole flow:
- A local AI model (LM Studio or Ollama) can take a couple of minutes for an hour-long meeting on a laptop.
- A cloud key (Anthropic, OpenAI, or Groq) turns that into seconds — only the transcript text is sent, never your audio, billed to your own key. See Configuring AI providers.
- No AI at all — use the copy-paste workflow and summarize in your own ChatGPT or Claude. See Running Daisy fully offline.
A cloud key speeds up summaries only, and never sees your audio.
If your machine feels slow
- Keep the built-in model. On Windows and Linux, the bundled Parakeet model is the fast path and needs no GPU. On a Mac, Apple's engine is. Switching to a large Whisper model trades a lot of speed for accuracy you may not need on clean meeting audio.
- If you do run Whisper, check the GPU is being used. Updating your graphics driver is the single biggest speed-up on a PC, because a machine with no working driver silently falls back to the CPU path.
- Hide the captions if the UI feels heavy. Settings → Recordings → Live captions is display-only; turning it off stops the recording view repainting constantly. The finished transcript is the same.
- Tune the engine, if you know what you're changing. Settings → Providers → Advanced → Transcription tuning exposes the window, hop, commit margin, and thread count per engine, each with a Reset to defaults button. The defaults suit most machines; this is there for the ones they don't.
- Bring a cloud key for fast summaries if the AI step is what feels slow.