Documentation Index
Fetch the complete documentation index at: https://docs.hyprcore.ai/llms.txt
Use this file to discover all available pages before exploring further.
Supported
- macOS 13 Ventura or newer. Apple Silicon (M1, M2, M3, M4) and Intel Macs are both supported. Apple Silicon is significantly faster for local Whisper models because Hyprcore uses Metal.
- 8 GB RAM minimum. 16 GB is comfortable for the larger models.
- 5 GB free disk space. Most of this is for speech models you choose to download. You can keep just one or two installed.
- A microphone. Built-in is fine for dictation; an external mic is better for long meetings.
Recommended
- Apple Silicon Mac with 16 GB RAM. All local models run smoothly, including Whisper Large.
- macOS 14 Sonoma or 15 Sequoia for the most reliable system audio capture in meetings.
What works on which hardware
| Model | Apple Silicon | Intel Mac | Notes |
|---|---|---|---|
| Parakeet V3 (CPU) | Excellent | Excellent | ~5× real-time on a recent Mac. Best default for dictation. |
| Whisper Small | Good | Good | Fast, decent accuracy. |
| Whisper Medium | Good | OK | Good balance for meetings. |
| Whisper Turbo | Excellent | Slow | Use on Apple Silicon if you want speed on long audio. |
| Whisper Large | Excellent | Slow | Highest local accuracy. |
Network
Hyprcore needs internet access to sign in and to fetch your speech models on first launch. After that, transcription itself runs entirely on your Mac. Network is also used for:- Auto-updates.
- Cloud transcription, AI summaries, and sync (only if you turn them on).
- The web dashboard at app.hyprcore.ai.