Support

We try to answer everything within a few days.

Get help

The fastest ways to get help with Vays:

  1. Check the FAQ on the main page. Most common questions are answered there.
  2. Join the Discord community at vays.app/discord. It is the quickest channel, other users can chime in, and answers stay searchable for the next person.
  3. Email [email protected] if your question contains private information you don't want to post publicly.

Reporting a bug

Vays keeps a privacy-safe Activity Log on your device: the last steps the app took, with no transcript text, audio, or note content, ever. Attaching it turns "it doesn't work" into something we can actually fix.

Along with the log, please include:

On iPhone there is also a Device check in Settings: a speed test of the whole transcription pipeline on your phone, useful as a first diagnostic when things feel slow.

Purchases & refunds

There is no subscription in Vays. The iPhone and iPad app is free, with an optional tip jar that unlocks nothing. The Mac app is a one-time purchase on the Mac App Store. All payments, receipts and refunds are handled by Apple, not by us:

Common issues

"My iPhone can't find my Mac"

Make sure both devices are on the same Wi-Fi network and that the Mac app is running: the sidebar status footer should read "Ready to pair" or "N devices paired"; "Sync server off" or "hidden" means the Mac is not discoverable. On the iPhone, open Settings → Privacy & Security → Local Network and confirm Vays has permission. If the Mac is on a corporate or guest network that blocks Bonjour, sync won't work; switch to a normal home Wi-Fi.

"On-device transcription failed"

The first transcription warms up the model and can take 10-30 seconds; subsequent transcriptions are much faster. If it fails, go to Settings → Transcription, delete the model, and re-download it. Make sure you have at least 1 GB of free disk space. The Settings screen also shows an Active engine row, so you can see which model actually transcribes your language.

"Gemma chat says it can't load the model"

The Gemma model download needs about 3 GB of free storage (plus about 1 GB more if you enable image understanding), and the on-device AI works best on devices with 8 GB of RAM. On iPhones with less memory Vays guards the AI features and defaults to clean transcription, so a recording never dies mid-processing; close other apps and try again if the model fails to load. The chat uses a serial inference queue, so opening chat while a recording is auto-processing will wait for the recording to finish first.

"My Polish transcription is poor"

If you're using WhisperKit Small as the engine on iPhone, switch to Parakeet TDT v3 in Settings → Transcription → Engine. Parakeet is the recommended engine for European languages and is the default for new installs. For Japanese, Chinese, Korean and Hindi, Vays automatically selects the multilingual NVIDIA Nemotron model.

Privacy & data requests

Vays does not collect or store any of your data on our servers; we have no servers. There is therefore no data for us to delete, export, or transfer. If you want to delete your Vays data, simply delete the app from your device; iOS / macOS will remove the entire app sandbox including all your recordings, transcripts, summaries, and settings.

If you have a specific data request, contact [email protected].

Feature requests

Post your idea on Discord or email [email protected]. Please describe what you want, why you want it, and how you'd expect it to work.

Contact

Discord: vays.app/discord
Email: [email protected]
Website: vays.app