Home/Reviews/Muah.AI
MUA

Muah.AI Review 2026

★★☆☆☆ 2.5
Breach History

Real-time phone calls almost nobody else offers, wrapped around the worst security event in this category’s history. Read the September 2024 section before you decide anything else.

Visit Muah.AI
4.9
Overall Score
Verified vendor docs, or two or more independent sources
Reported one source — plausible, unverified in-product
Contested sources disagree; the disagreement is the finding

The Verdict

We are scoring Muah.AI low on trust, not on features, and we want to be explicit about that so the number is not mistaken for a review of the chat. It has a real technical distinction — live phone calls where the AI picks up and carries context from recent chats, which reviewers call the only true implementation in the category. It is also the product whose 2024 breach became the category’s defining trust event and reset how every reviewer here advises people to sign up.

The jank alongside the distinction is well documented: calls reportedly drop every two to three minutes, there is a rough silent warm-up, and one tester’s AI forgot his name mid-call. Reported And custom character creation is locked until account level 10 — a progression gate used as a retention device, and the only one of its kind we found. Reported

Best for

Almost nobody, at this score. If you want live AI phone calls specifically, it is one of very few options — go in with a throwaway email and a virtual card.

Skip it if

You would be harmed by your chat prompts and email address becoming public. That is not hypothetical here.

Category pages score each app against that category’s own criterion — a “Chat Score” and an “Image Score” measure different things. The number above is the overall score.

The September 2024 breach

This is verified by major outlets, not a rumour. On 17 September 2024, 1.9 million email addresses and chat prompts leaked. The hacker described the site as “open-source projects duct-taped together.” It was first reported by 404 Media, which found tens of thousands of prompts describing child sexual abuse among the data, with prompts linked to real-name personal email addresses. The data was subsequently used for extortion. Verified

The fallout: traffic reportedly fell to around 600–900K monthly visits from a roughly 2M-user 2024 peak; its subreddit was banned after a documented astroturfing and vote-manipulation campaign using aged bot accounts to downvote criticism; and it was flagged by the Australian eSafety Commissioner. Reported

The most useful thing about this incident is what it changed for everyone else. Reviewers across this entire category now advise throwaway emails and virtual cards as standard — and they do it because of this event. If you take one operational habit from this site, take that one.

What it actually does well

Real-time phone calls on the top tier: the AI answers and carries context from recent chats. Multiple reviewers describe it as the only platform doing genuine live calls. Reported The execution is rough — calls reportedly dropping every two to three minutes, a silent warm-up period, and at least one report of the AI forgetting the caller’s name mid-conversation.

The tiering is flat — no tokens, no credits — with memory depth itself sold as a tier feature (“expanded memory,” “max memory”). Free users get a degraded experience — low-resolution images and ads — rather than blurred content. Reported

Muah.AI Pricing

Snapshot: July 2026 · prices in this category change quarterly — check the live page before paying
TierPriceWhat it buys
Free$0Degraded rather than blurred — low-resolution images, ads
VIP$19.99/mo ($69.99/yr)Base paid tier. Reportedly crept up from an earlier $9.99
UHD VIP$49.99/moHigher-fidelity images, expanded memory
ULTRA VIP$99.99/moReal-time phone calls, maximum memory

Flat tiers with no token meter, which is a genuine plus. Note that memory depth is itself a paid tier feature — the thing users complain most about across this whole category is here sold as an upgrade.

What Reviewers Say vs What Users Say

Across this category, editorial ratings and user ratings are close to inversely correlated. We publish both, because the gap is usually the most useful number on the page.

Editorial / affiliate sites
Mixed
Reviewers credit the calls, and lead with the breach
Actual users
Damaged
Subreddit banned after a documented astroturfing campaign

Muah is the one product here where the editorial layer and the user layer agree, because the breach is not a matter of opinion. The vote-manipulation campaign that got its subreddit banned also means ordinary community sentiment is hard to read — some of what is out there was manufactured. Reported

What people praise
  • Genuine real-time phone calls, which almost nothing else offers
  • Flat pricing with no token meter
  • Free tier is degraded rather than paywalled shut
What people hate
  • The September 2024 breach — 1.9M emails plus chat prompts, then extortion
  • Calls reportedly drop every two to three minutes
  • Character creation locked until account level 10
  • Memory depth sold as a paid tier
  • Base price crept from around $9.99 to $19.99

Incidents & Controversies, Dated

Products in this category change under their users constantly. These are the events worth knowing about before you commit a character — or a year’s subscription — to one.

17 Sep 2024 — The breach Verified

1.9 million email addresses and chat prompts leaked. The hacker described the site as “open-source projects duct-taped together.” 404 Media reported tens of thousands of prompts describing child sexual abuse in the data, with prompts linked to real-name personal emails. The data was subsequently used for extortion. Reported by 404 Media, Linklaters and Malwarebytes among others.

Post-2024 — Traffic collapse, subreddit ban, regulator flag Reported

Traffic reportedly fell to around 600–900K monthly visits from a ~2M-user peak. The subreddit was banned after a documented astroturfing and vote-manipulation campaign using aged bot accounts to downvote criticism. Flagged by the Australian eSafety Commissioner.

Muah.AI FAQ

Is Muah.AI safe?

It had the worst security event in this category’s history. In September 2024, 1.9 million email addresses and chat prompts leaked, prompts were linked to real-name personal emails, and the data was subsequently used for extortion. If you use it, use a throwaway email and a virtual card — advice that now applies across this whole category largely because of this incident.

Does Muah.AI really do phone calls?

Yes, on the $99.99/mo ULTRA VIP tier, and reviewers describe it as the only platform doing genuine live calls where the AI picks up and carries context from recent chats. The execution is rough: reports of calls dropping every two to three minutes, a silent warm-up, and at least one case of the AI forgetting the caller’s name mid-call.

Why can’t I create a character on Muah.AI?

Custom character creation is reportedly locked until your account reaches level 10, so you have to engage with existing characters first. It is an explicit progression gate used as a retention device, and it is the only one of its kind we found in this category.

How much is Muah.AI?

VIP is $19.99/mo ($69.99/yr), UHD VIP is $49.99/mo and ULTRA VIP — the tier with the phone calls — is $99.99/mo. Flat tiers, no tokens. The base price reportedly crept up from an earlier $9.99, and memory depth is itself one of the things a higher tier buys.

Where Muah.AI Ranks

How This Review Was Built

Compiled from vendor documentation and independent review sources during a read-only research pass in July 2026 — not from a paid subscription run on every product. Every claim below carries the confidence marker it arrived with, and anything we could not confirm was left out rather than asserted. Full method.

We may earn a commission if you sign up through a link on this page. That has no effect on the score, and the score above was set before any affiliate relationship was considered. We do not own or operate Muah.AI; for the four sites in this market that we do own, see our ownership disclosure.

Primary sources

https://www.404media.co/hacked-ai-girlfriend-data-shows-prompts-describing-child-sexual-abuse-2/
https://www.linklaters.com/en/insights/blogs/digilinks/2024/october/the-muah-ai-data-breach---extortion-threats-and-cyber-vulnerabilities
https://www.malwarebytes.com/blog/news/2024/10/ai-girlfriend-site-breached-user-fantasies-stolen
https://plisio.net/ai/muah-ai