# FiltreAI — On-Device SMS Protection

> On-Device SMS Protection (iPhone/iPad app by EMIRHAN KESKIN.)

- Source: https://appshunter.io/ios/app/filtreai/id6797799385 (this page in markdown: same URL + `.md`)
- Developer: [EMIRHAN KESKIN](https://appshunter.io/developer/6797799387)
- Category: Utilities
- Price: Free
- Age rating: 4+
- Requires: iOS 15.1 · 34 MB
- Languages: American English
- Released: 2026-08-14
- Data updated: 2026-10-08

## What is FiltreAI?

FiltreAI helps you classify junk, promotional, transactional, and suspicious SMS from unknown senders directly on your device.

Why FiltreAI?
• Automatic on-device classification with the iOS Message Filter extension
• Sender, content, regular-expression, and known suspicious-domain rules
• Junk, transaction, and promotion categories
• An on-device allowlist for senders you trust
• Smart Analysis to assess a message before you act
• A Test Simulator for previewing a sample classification without sending an SMS
• A Messages reporting panel that can create a local sender rule for future messages
• A built-in protection pack that works without internet access

Privacy comes first
Filtering and Smart Analysis run on your device. The FiltreAI developer does not collect messages, senders, contacts, device identifiers, or usage data from this app. There is no account, advertising, analytics SDK, remote diagnostics, or server notification service. The threat list is included in the app and updated through new App Store versions.

How it works
Enable FiltreAI in the Message Filtering section of iOS Settings. When iOS offers an eligible message from an unknown sender to the extension, FiltreAI evaluates the built-in protection pack, your rules, and your allowlist on the device. Messages that do not match are left allowed.

FiltreAI is a supporting security tool. It does not delete messages and cannot guarantee error-free results. Verify the sender before opening a suspicious link or sharing personal information.

## Is FiltreAI safe?

Not enough ratings yet to assess FiltreAI.

## Is FiltreAI legit?

FiltreAI has limited verification signals. The AppsHunter Legitimacy Score for FiltreAI is 58 out of 100 points across 7 verification signals from App Store and developer data.


## Version history (last 5 releases)

### 1.2.6 — 2026-10-01

FiltreAI 1.2.6 — stronger betting filter, plus an explanation on iPhone:

• Betting messages using the "wagering-free bonus" phrasing are now caught. The phrase is betting jargon; a user reported one and we measured it: the message reached the inbox without matching a single warning signal.
• The same rule now also recognises the spelling without Turkish diacritics, which is how these messages are commonly written and which previously fell outside the rule.
• One more link shortener was added to the recognised list.
• On iPhone, the Home screen now explains why the counters and Recent Activity stay empty: for privacy, Apple does not report the filter's decisions back to FiltreAI. Filtering continues; an empty list is not a fault.

Legitimate messages are unaffected: bank bonus notices and campaign texts keep coming through exactly as before.

Filtering keeps running on your device: no account, no internet connection, no data collection.

### 1.2.5 — 2026-09-06

FiltreAI 1.2.5 — filtering and interface fixes:

• Companies that write their trader registration number differently are now recognised: the "Mersis Numarası:" wording and numbers grouped with hyphens could not be read. Insurance and hospital messages using those forms stay in Promotions instead of Junk.
• Operator code warning: codes like B001 or B002 at the end of a message identify the operator that delivered it, not the sender, and almost every legitimate bulk SMS carries one. Such a code can no longer be added as a rule, and existing rules are flagged in the list.
• Messages tried in the Test Simulator are no longer written to Recent Activity. What you type there is not a message you received; the result still appears on the simulator screen.

Gambling, fraud and phishing filters keep running exactly as strictly as before.

Filtering keeps running on your device: no account, no internet connection, no data collection.

### 1.2.4 — 2026-09-01

What's new in FiltreAI 1.2.4:

• A brand's own short domain is no longer treated as junk: invoice, ticket and campaign messages were being blocked purely because of the link's extension. The measured cases were an operator invoice reminder, a bus ticket notice and a bank campaign.
• The known-shortener list works correctly on the device again: it was applied as a plain text search, so any message merely containing a shortener name — including corporate campaigns carrying a trader registration number — went straight to Junk. A shortener now only feeds the link assessment and never decides on its own.
• The iOS message filter extension recognises known shorteners even before the app is first opened.

Gambling, fraud and phishing filters keep running exactly as strictly as before.

Filtering keeps running on your device: no account, no internet connection, no data collection.

### 1.2.3 — 2026-08-28

What’s new in FiltreAI 1.2.3:

• Reporting from the Messages panel works again: “Start Report” now opens FiltreAI with the sender and message ready to confirm. In the previous release this on-device transfer could not complete.
• Short-form business identity is recognised: the “M:” trader registration number common in Turkish commercial messages is now read, so senders using that format stay in Promotions instead of Junk.
• Changing the category in the panel and reporting again no longer leaves the earlier draft in pending reports.
• When the transfer cannot complete, the app now says so plainly.

Filtering keeps running on your device: no account, no internet connection, no data collection.

### 1.2.2 — 2026-08-22

What's new in FiltreAI 1.2.2:

• Unregistered bulk sender detection: Senders that carry commercial message markers but no verifiable business identity are now classified as junk.
• Registered business identity: A trader registration number in the message body is recognised, so legitimate corporate senders stay in the promotion category.
• One-tap setup: The setup step in Settings now opens the iOS Message Filtering screen directly.
• Protection status is now evidence-based: Whether the filter is running is confirmed from records the extension produced on device, replacing the previous assumed "active" state.
• When an app icon change does not apply, the app now says so instead of failing silently.
• Screen reader labels and interface refinements.

Filtering still runs entirely on your device — no account, no internet connection, no data collection.

## Apps similar to FiltreAI

| App | Rating | Price | Category |
| --- | --- | --- | --- |
| [Kivi Printer](https://appshunter.io/ios/app/kivi-printer/id6478200680) | — | Free | Utilities |
| [GPS Phone Locator: FamLocate](https://appshunter.io/ios/app/gps-phone-locator-famlocate/id6723885198) | — | Free | Utilities |
| [SUNNY HERO](https://appshunter.io/ios/app/sunny-hero/id1576713289) | — | Free | Utilities |
| [ArkSigner](https://appshunter.io/ios/app/arksigner/id1496600667) | 1.0 (1) | Free | Utilities |
| [iledeyes](https://appshunter.io/ios/app/iledeyes/id6504541936) | 3.2 (5) | Free | Utilities |
| [Locklizard Safeguard Viewer](https://appshunter.io/ios/app/locklizard-safeguard-viewer/id627385747) | 1.8 (8) | Free | Utilities |

---

*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 2026-10-08.*
