Is the Ascension App Accurate? An Honest Review of the Facial Analysis App

You upload the front photo, add the side profile, finish the onboarding, and wait for the promised scan. Instead of seeing your tier or percentile, you meet the payment screen. If you pay, you may then get a polished dashboard of angles, rankings, and recommendations — and a charge that is harder to interpret than the face map.

Ascension measures real facial angles with a standard landmark model, then turns them into a tier and percentile scored against a “human consensus” it does not publish — and its own policy says it does not guarantee accuracy. Every core result is paid, and despite the “photos never leave your phone” pitch, its privacy policy says the front photo goes to Google Cloud for the rating.

Desk research as of October 7, 2026; we bought nothing.

Key numbers

• 4.77874 average from 5,121 ratings on the U.S. App Store listing. Apple App Store
• 4.5 stars, 3.35K reviews, and 50K+ downloads on Google Play. Google Play
• $4.99 a week is the listed weekly Pro price; $49.99 is the listed lifetime price. App Store listing
• 478 facial landmarks are described in the privacy policy for front and side-profile photographs. Ascension privacy policy
• Version 4.4.2 was updated on September 26, 2026. Apple App Store

What does the Ascension app actually measure?

Ascension starts with face landmarks, not a mystical overall impression: its policy describes 478 normalized x, y, and z coordinates from front and side photographs, while the store names canthal tilt, gonial angle, maxilla projection, facial thirds, cheekbone-to-jaw ratio, and symmetry. Supported devices also get a TrueDepth/ARKit 3D mesh and depth-based heatmap.

Landmarks offer a repeatable ruler, but interpretation is less concrete: the store says measurements are scored against human consensus, while the homepage gives no rater count, dataset, or calibration method; it lists 14 signals versus the store’s 28 “clinical-grade metrics.”

A measurable angle does not validate the ranking. See our reviews of Areum and the PSL looksmax app.

Is the Ascension app accurate? (Ruler vs jury)

Ascension appears to measure landmarks, but accuracy for its tier or percentile is not established. We found no published paper, benchmark, test-retest study, rater-dataset size, or validation result; its privacy policy and terms warn results may not be accurate or reliable.

Ruler versus jury is the model: the ruler is landmark geometry; the jury is the undisclosed consensus deciding what those numbers mean. Updates labelled “Rating Model Improvements” and “Improved Model Accuracy” make scores from different versions non-comparable — an inference from version history.

One 3-star U.S. reviewer wrote on March 28, 2026 that scans changed from the 80th percentile to the 20th and then 65th.

For face-rating claims versus validation, see whether face-rating apps work. For score instability across apps, see why face-rating apps give different scores.

One user's experience is not a test, but a published test-retest study would have to rule it out.

Photo: Yan Krukau on Pexels

Is Ascension free? Where does the paywall appear?

Ascension is free to download, but the core result is gated: the store puts the rating, metrics, stack, 3D scan, and AI Coach under Ascension Pro, and users report hitting the paywall after uploading photos, before seeing the result.

A U.S. 1-star review on June 10, 2026 said, “Everything up until you subscribe in the intro is nice and polished ... then it'll ask you to pay”; a Google Play reviewer wrote on July 4, 2026, “when you finally click for it to evaluate, you get slammed into a paywall.”

Reviews mention inviting three friends, referral-code searches, and a Creator Program. Exact terms, limits, and eligibility were not published; do not treat “free code” as guaranteed. Our paywall-after-upload explainer explains why placement matters.

How much does Ascension cost over a year?

The price sheet is messy: multiple yearly prices and both “Pro Access” and “Ascension” labels appear in the U.S. App Store listing. The cheapest apparent route is $19.99 yearly; weekly becomes expensive over a full year, while Google Play shows $7.99–$39.99 per item.

Three different “Pro Access (1 Year)” prices suggest discounting or A/B-tested paywalls, though that is an inference. One 5-star U.S. reviewer wrote on January 21, 2026: “It only cost 20$ for a YEARLY PAYMENT,” which fits the lowest listing but is not necessarily every user’s offer.

What do Ascension reviews actually say?

Public reviews are mixed, not a clean survey: the U.S. most-recent RSS feed returned no entries, so this sample is the 10 Apple-selected reviews; we also used 17 U.K. RSS reviews and three Google Play reviews.

The Apple-selected U.S. breakdown was three 5-star, one 4-star, four 3-star, and two 1-star reviews. A 5-star reviewer on December 30, 2025 called the scan “vry detailed”; a 4-star reviewer on September 3, 2026 said the app “rates you and goes into detail” after purchase.

Criticism focuses on stability, advice, and the paywall. A 3-star reviewer on February 25, 2026 reported “a heavily recessed maxilla” despite “my forward growth is one of my strengths,” plus a liposuction recommendation. A 1-star reviewer on September 12, 2026 said the result was paywalled and “super random, nothing meaningful.”

The U.K. sample was 10 five-star and seven one-star reviews, with no 2–4-star entries. A 5-star review on August 18, 2026 said it “Does the job better than any fake looksmaxximg time wasting apps”; a 1-star review on September 5 called it “not worth the money at all.”

These samples cannot estimate every user’s experience or validate the model.

Is Ascension safe? Where do your photos go?

The policy answer is that front-profile photos leave the device: encoded images go to Google Cloud Platform for AI rating analysis and are processed in real time. That conflicts with the store and homepage promise that scans run on-device and photos never leave the phone.

The iOS privacy label says Usage Data tracks users, declares usage data and identifiers for third-party advertising, and omits a Photos/Videos category. Google Play says the app may share Photos and videos and device or other IDs with third parties, collects personal information, encrypts data in transit, and allows deletion. See our face-rating app privacy audit for that mismatch.

The policy keeps 3D meshes, depth point clouds, landmarks, and side-profile photos on-device; OpenAI receives derived metrics, not raw photos or depth data. It names GCP, OpenAI, Mixpanel, Superwall, RevenueCat, and Apple as processors. No explicit clause says photos train models; the homepage’s distribution claim leaves the long-term story unclear.

There is an age mismatch: the policy says 18, the terms 13, and the App Store 9+. The homepage recommends peptides and drugs including MK-677 and Retatrutide; those need a doctor, not casual app tips.

The policy’s on-device storage claim for 3D data does not erase the front-photo upload contradiction.

How do you cancel Ascension and get a refund?

On iPhone, open Settings, tap your name, choose Subscriptions, select Ascension, and cancel. Apple says cancellation should happen at least 24 hours before renewal; deleting the app does not cancel it. Ascension’s terms send refund requests to Apple, not Ascension.

For an Apple refund, use reportaproblem.apple.com, choose “I’d like to,” then “Request a refund.” Android users must cancel through Google Play subscriptions; its refund route was linked on the listing, but we did not verify Google’s support page. For comparison, see how to cancel a looksmax subscription.

Is Ascension worth it, and what's the free alternative?

Ascension is worth considering only for paid access to a detailed landmark and angle dashboard. It is not a strong value if you expect a free result, a published consensus dataset, stable percentiles, or a clear boundary between measurement and subjective judgment.

The better question is not “what tier am I?” but “what first impression do people receive?” Our free first-impression test reads your whole first impression (face, physique, outfit). The read and breakdown are free on a free account with no card; deeper modules are $1.50 each, one-time, never a subscription. It is a perception test, not a clinical instrument; no photo-based test is.

That is a different promise, not objective truth. If rankings make you spiral, pause before buying another number. A useful read should reduce pressure, not turn every mirror check into a verdict. Our comparison of free looksmaxxing apps with no paywall lays out an alternative.

The bottom line

Ascension has a real measurement layer: facial landmarks, angles, a 3D scan path, and a sizeable public store presence. But its decisive output — the tier and percentile — rests on a human-consensus model whose dataset and validation are not published, while the app’s own policy disclaims accuracy. You pay before seeing the core result, and the photo-privacy pitch does not match the policy.

So our verdict is no: Ascension is not accurate enough to treat its percentile as a fact about your attractiveness. It may be useful as a paid geometry dashboard if you understand the limits, verify the subscription price before confirming, and accept that the “jury” is hidden. For the perception question, try our free first-impression test before committing money.

A face can affect how people respond to you, and that first impression can form in about a tenth of a second — but that effect is more movable than any app’s number. Willis and Todorov’s 2006 study is about rapid judgments from a face image, not a life sentence. That is the useful frame: understand the signal, then work on what you can actually change.

Sources

• Ascension - Facial Analysis on the U.S. App Store
• Ascension - Facial Analysis on Google Play
• Ascension privacy policy
• Ascension terms of service
• Ascension homepage
• Apple: cancel a subscription
• Apple: request a refund
• Google MediaPipe Face Mesh documentation
• Willis & Todorov (2006), First impressions

Prices, ratings, privacy labels and reviews as shown in publicly available App Store (and Google Play, if applicable) listings and user reviews (U.S., October 7, 2026).