Support
We read every message. Replies usually within a day.
Contact us
Email essentialdevteam@gmail.com. Please include your iPhone model, iOS version, and the model you were using (Settings › About shows the app version) so we can help faster.
Report a bug or request a feature
In the app: Settings › About › Report a bug / Request a feature. Both open a pre-addressed email.
Frequently asked questions
Which model should I download first?
Open Models › Fits. The first row is marked Best for you — it's the highest-quality general model that runs comfortably on your iPhone. On most recent iPhones that's a 4B model such as Gemma 3 4B or Qwen3 4B.
What do "Fits well", "Fits", "Tight" and "Too big" mean?
They compare the model's memory need (weights + context) against what iOS lets an app use on your phone. Fits well uses under 55% of that limit and stays fast. Fits works well; very long chats may slow down. Tight will run but close other apps first. Too big can't load, so there's no download button. Tap ⋯ › How fit is calculated in Models for the full explanation.
Does it work without internet?
Yes. Internet is needed only to download a model. After that, everything — chatting, thinking models, photos with Gemma 3 — works in airplane mode.
The download stopped or failed.
Downloads continue in the background and survive closing the app. If one stops, open Settings › Storage: you'll see Pause / Resume / Retry and, under ⋯, Restart from zero. If the file fails its integrity check, the app deletes it and retries once automatically. Downloads are Wi-Fi-only by default; change that in Settings › Storage.
Why does the first reply take longer?
The model is loaded into memory on first use, and the first time a model runs on your phone the graphics shaders are compiled. Later replies start in well under a second. Turn on Keep model loaded in background in Settings › Generation if you switch apps often.
What is "Thinking"?
Reasoning models (DeepSeek R1, Qwen3) work through the problem before answering. QUIET shows that as a collapsible "thinking" note above the reply. In Settings › Generation you can show it, hide it but keep it, or discard it. You can also turn thinking off for Qwen3 for faster answers.
The app said it freed memory. What happened?
iOS asked for memory back (usually because other apps or Safari tabs are open). QUIET unloaded the model instead of crashing and kept your chat. Tap Reload, or close other apps and try a model with a better fit rating.
Can I add a model that isn't in the list?
Yes, with QUIET Ai Pro: Models › ⋯ › Add from Hugging Face. Paste a public repository like bartowski/…-GGUF, pick a file, and the app reads its header to check fit before downloading. Any public GGUF works, including community fine-tunes. Split files and gated repositories aren't supported.
What does Pro include, and how is it priced?
Pro unlocks any number of downloaded models, Hugging Face import, and scheduled iCloud Drive backups. Choose monthly ($2.99), yearly ($19.99, with a 7-day free trial) or lifetime ($29.99 once, no subscription). Subscriptions renew automatically unless cancelled 24 hours before the period ends — manage them in Settings › your name › Subscriptions on your iPhone. To restore on a new phone: Settings › App › Restore purchases.
How do I back up or move my chats?
Settings › Backup › Export all chats makes an encrypted backup file you can save anywhere; Restore from backup reads it on another phone. Chats are also part of your normal iPhone backup unless you turn that off. Models are never backed up — they're re-downloaded.
How do I delete everything?
Settings › Danger zone: Delete all chats, Delete all models, or Reset app. Deleting the app also removes everything, because nothing is stored outside your phone.