If something isn’t working, start at the top of this list. Most problems fall into one of these categories.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.
Dictation issues
Text doesn’t appear after I speak
Almost always a missing Accessibility permission.Open System Settings → Privacy & Security → Accessibility
Confirm Hyprcore is in the list and the toggle is on.
Shortcut doesn’t trigger
Check Input Monitoring permission in System Settings → Privacy & Security. Same restart-after-grant rule. If permissions are right but the shortcut still doesn’t fire, try:- A different shortcut. Some combinations are intercepted by macOS, BetterTouchTool, Karabiner, or other shortcut managers.
- Switch the keyboard implementation in Settings → Advanced → Keyboard implementation between HandyKeys and Tauri. Restart required. HandyKeys is more robust on most systems.
Audio looks empty (waveform doesn’t move)
- Check Microphone permission.
- Check Settings → Audio → Input device—you may have an unused device selected.
- Check macOS isn’t muting Hyprcore: System Settings → Sound → Input while talking. The level meter should move there too.
Transcription is wildly wrong
- Try a larger model. Whisper Small → Whisper Medium → Whisper Turbo or Large. Larger models are dramatically more accurate.
- Set the right language in Settings → Models → Speech model → Language.
- Reduce background noise; Whisper handles babble worse than silence.
- For technical jargon, try post-processing with a “fix proper nouns” template that lists your terms.
Meeting issues
”Screen recording permission required”
Open System Settings → Privacy & Security → Screen & System Audio Recording. Add Hyprcore. Restart Hyprcore.System audio isn’t being captured
macOS needs System Audio permission specifically (a sub-permission of Screen Recording on macOS 14.4+). Confirm in System Settings → Privacy & Security → Screen & System Audio Recording. If that’s set and audio still isn’t captured, the call app might be using exclusive audio mode. Restart the call app and try again.Recording is laggy
- Lower Video quality in Settings → Meetings (try Medium or Low).
- Switch to a smaller speech model—Whisper Medium is plenty for most meetings.
- Use Parakeet V3 if your Mac is CPU-only; it’s faster than Whisper.
- Check Activity Monitor. If
hyprcoreis using >100% CPU continuously, the model is too heavy for your hardware.
Live transcript is delayed
Local models batch every few seconds. If you want sub-second live text, switch to a cloud STT provider (Settings → Meetings → STT provider → Deepgram or Groq).Model and download issues
Model download fails
Ifmodels.hyprcore.ai is blocked by a firewall or proxy, do a manual install.
If downloads fail mid-way, delete the partial file from ~/Library/Application Support/ai.hyprcore.desktop/models/ and try again.
Whisper Large or Turbo crashes the app
Known issue on some configurations. Workarounds:- Switch to Whisper Medium. Quality is similar; stability is much better.
- Use Parakeet V3 if you don’t need Whisper specifically.
- On Intel Macs, larger Whisper models stress the CPU—consider cloud STT for long meetings.
”Model not found” after manual install
- Filename or directory name has to match exactly. See Manual install.
- Restart Hyprcore after copying files.
- For Parakeet, the directory name is
parakeet-tdt-0.6b-v3-int8—not the archive name.
Sync and account issues
”Sign in failed” / browser doesn’t return to the app
- Check that the deep-link
hyprcore://is registered. Quit Hyprcore, reopen it, and sign in again—the deep link is registered on launch. - If your browser shows the auth page but doesn’t redirect: copy the link manually from the browser address bar and paste into the desktop sign-in screen.
Sync seems stuck
- Settings → Sync → Test connection. If it fails, the issue is network.
- Settings → Sync → Force sync. Re-runs the comparison and pushes/pulls anything missing.
- Restart Hyprcore. Sync is resilient but a full restart fixes most transient issues.
My credits ran out unexpectedly
- Open app.hyprcore.ai/dashboard/usage for the breakdown.
- Cloud STT is the most credit-heavy operation. Consider switching to local Whisper for routine recording.
- Bring your own API key for the provider you’re using most. Hyprcore Cloud has no markup, but BYOK avoids credit usage entirely.
App won’t start / crashes
”Hyprcore quit unexpectedly” on launch
- Open Console.app and check the crash log under “Crash Reports”.
- Try launching with
--debug: - If a recent setting change broke things, reset settings: rename
~/Library/Application Support/ai.hyprcore.desktop/settings.dband relaunch (Hyprcore creates a fresh one). - As a last resort,
rm -rf ~/Library/Application\ Support/ai.hyprcore.desktop/to fully reset—you’ll lose dictation history and local settings, but cloud-synced data comes back when you sign in.
High CPU when idle
Make sure Always-on microphone is off (Settings → Audio). When on, Hyprcore keeps the mic stream open, which uses some CPU continuously.Still stuck
- Live chat. The chat bubble on hyprcore.ai and inside the web dashboard. Fastest way to reach the team in real time.
- Discord. discord.gg/UEvrjftB—other users have probably hit the same issue. Include your debug logs (
Settings → Advanced → Open log directory). - Email. support@hyprcore.ai for billing, account, or anything you’d rather not post publicly.
- In-app feedback. Help → Send feedback inside Hyprcore attaches your version and recent logs automatically.

