# AiWraith — Private on-device AI

> Private on-device AI (iPhone/iPad app by Sheng Gao.)

- Source: https://appshunter.io/ios/app/aiwraith/id6797142575 (this page in markdown: same URL + `.md`)
- Developer: [Sheng Gao](https://appshunter.io/developer/6797142577)
- Category: Productivity, Utilities
- Price: $0.99
- Age rating: 17+
- Requires: iOS 17.0 · 41 MB
- Languages: American English
- Released: 2026-08-19
- Data updated: 2026-09-04
- User reviews in markdown: https://appshunter.io/ios/app/aiwraith/id6797142575/reviews.md

## What is AiWraith?

Wraith is a private AI assistant that runs entirely on your device. There's no account to create, no cloud to send your data to, and no tracking — everything happens locally on your iPhone.

Ask questions, hold a conversation, and get things done with an assistant that keeps your data where it belongs: with you.

WHAT IT DOES
• Chat with a capable on-device assistant — no login, no subscription wall
• Read and summarize documents you attach
• Transcribe audio on-device
• Capture tasks, plan your day, and get a daily brief
• Search the web only when you ask it to

PRIVATE BY DESIGN
• Runs open-source models directly on your device
• Nothing is collected, and nothing leaves your device unless you explicitly trigger a web search
• A No-Egress mode locks the app fully on-device — an airplane mode for your data
• See exactly what has left your device, and where, in the privacy ledger

WORKS WITH YOUR MAC (OPTIONAL)
Pair Wraith with a Mac on your local network to run larger models over Wi-Fi. It's an optional unlock — the app is fully useful on its own, offline.

Wraith requires no account, no sign-in, and no personal information to use.


## Version history (last 3 releases)

### 1.0.2 — 2026-08-23

Fixes the first thing you see.

The app runs its AI model on your iPhone, so it has to download that model once before it can answer anything. It used to start that download silently, behind your first question — so a normal one-to-three-minute wait looked exactly like the app had frozen.

Now it tells you. The opening screen asks you to download the model and says how big it is, the status reads "No model yet" until it's there, and the download shows a live percentage and elapsed time. If your connection drops it says so, and nothing is lost — asking again picks up where it left off.

You'll also see the data-sharing screen once more. It now names the model download alongside everything else, so the list of what leaves your device is complete.

### 1.0.1 — 2026-08-19

Model downloads now show real progress. The progress bar used to sit
at a few percent for minutes on a perfectly healthy download, then jump
to 100% — it now moves with the bytes actually received.

Loading a model into memory is now its own labelled step, so picking a
model you've already downloaded no longer shows a full, motionless bar
while it loads.

The recommended on-device model is now Qwen3.5 0.8B.

### 1.0 — 2026-08-19

No release notes.

---

*Data collected daily from the US App Store and indexed by [AppsHunter](https://appshunter.io/). User reviews are verbatim App Store reviews. Ratings, prices and chart positions refresh continuously; this snapshot is from 2026-09-04.*