Best ai photo apps

Background removal is fast and precise, and the library of fonts and backgrounds is extensive.

Glitches during batch processing of more than five photos, and AI outpainting produces artifacts.

Small businesses and marketplace sellers who need to create design assets.

Instantly creates presentations, posts, and logos from templates with no design skills required.

AI generation disappoints experienced users, and export is unreliable.

Content creators, students, and small businesses with no design background.

Precise sliders for skin, face shape, and body contour give real control over the degree of change. Makeup looks authentic. Acne removal and skin smoothing work cleanly.

Heavy editing makes results look plastic and unnatural. Editing multiple people in one photo applies changes only to the first face. Some users noticed texture quality declined after a late-2024 update.

People who want to lightly retouch a real photo for social media without building a synthetic look from scratch.

Wide selection of clothes, accessories, and face types. The ability to move and scale individual elements gives flexibility in composition. The absence of intrusive banners sets it apart from competitors.

Many elements do not support color changes: a chosen hairstyle comes in only one shade. Few skin tone and mouth expression options. Occasionally the character disappears after closing the app.

Teens and anime fans who want a character-building tool for messages, avatars, or stickers.

A live painting process with instant response to parameter changes, and an infinite number of unique results from a single photo.

Rendering has slowed noticeably on older devices, and white patches sometimes appear on unpainted areas.

Artists and photographers who enjoy a controlled creative process rather than a one-shot filter.

High likeness when 15-20 varied photos are uploaded, a wide selection of historical and fantasy styles, and headshots that look professional and get used for real tasks.

About half the shots in a set cut off the top or bottom of the head. Eyes sometimes look unnatural with misaligned pupils or mismatched layers. Over time the style catalog stops refreshing.

People who want quality avatars with strong likeness for a profile, presentation, or just for fun trying different eras.

Real improvement in video clarity without changing the content or adding a watermark. Noticeably improves export quality for editors working in other apps.

Photo enhancement sometimes shifts faces toward more Asian-looking features. The QHD function frequently throws a network error and fails to complete. The number of base-tier attempts has been cut with each update.

Video editors who want to improve final export quality without a complex pipeline.

The ability to upload a custom video for animation sets it apart from competitors. Lip sync is convincing and the result is genuinely surprising. Ease of use is praised by nearly everyone.

In the free version the video is blurry. High-quality results are sent by email with a delay or do not arrive at all. The head sometimes distorts when placed on templates with sharp angles.

People who want to bring family photos to life or create fun video greetings for loved ones.

Real-time filter intensity adjustment gives control over the result. A wide range of styles from watercolor to oil painting. The option to send a postcard directly from the app.

Elderly faces come out with exaggerated wrinkles. The watermark does not disappear for some users even after a purchase. Saving sometimes freezes and the image is lost.

Art enthusiasts and photographers who want a quick painterly effect without Photoshop.

Background protection during body work, which users call unique. Object removal (poles, signs, entire cars) is fast and clean. Body tools allow subtle adjustments.

After a recent update the AI recognizes faces less accurately and produces unnatural results. Muscle correction settings were simplified from three sliders to three fixed options, reducing flexibility.

People who regularly edit photos for social media and need quick background object removal or light body correction.

Facial tracking quality is good enough that the avatar looks alive during a stream. Extensive avatar appearance customization. A low barrier to entry for beginner VTubers without expensive equipment.

Face tracking has degraded compared to earlier versions, and the head sometimes locks into the wrong position. Full-body tracking is absent, only head and facial expressions. Several users report sudden permanent bans without warning and no appeal process.

Young people who want to try virtual streaming with an anime look without expensive equipment. Not suitable for children without adult supervision due to platform risks.

Keeps your identity intact: it is you, just at your best. Works well on group photos where someone is not looking at the camera. Natural retouching without a plastic look.

Eyes often look unnatural or glassy. Skin is sometimes smoothed to the point of losing texture. Results are unpredictable and there is no intensity control.

People who want to fix one specific moment in a photo without radically changing their appearance.

Makeup overlay looks natural and not overdone on most photos. A large choice of colors and styles. Long-term users since 2016 note stable core quality up until recent updates.

The app automatically whitens teeth and smooths skin with no way to turn it off, making people look unrecognizable, especially those who already have light teeth or freckles. Some features work poorly with glasses or deeper skin tones. After recent updates some filters stopped loading.

People who want to try on makeup before buying or before going out. Also suitable for those who want to see themselves with a different hair color or eye style in a photo.

When it works, the result looks like a real professional shot with correct facial features and lighting. One user wrote it was the first AI-generated photo that actually looked like them.

Appearance is unpredictable: different hair color, different body type, and dark-skinned users see blotches and skin lightening. Several users received images in underwear instead of business attire. The result cannot be previewed before receiving it.

People who need a quick business portrait for LinkedIn and are comfortable with unpredictable quality.

The cartoon character is recognizable and resembles the face in the photo when the source image is good. Friendly interface, animated stickers with expressive facial gestures.

Limited clothing selection, especially for male characters. The avatar cannot be used directly in messengers. The beard and some clothing items produce layering artifacts.

Teenagers and parents who want a safe cartoon avatar for messengers and profiles.

A rich collection of backgrounds and stickers, and portrait retouching for experienced users.

Avatars are disproportionate and do not resemble you, and the output is hard to predict.

Collage and sticker users who do not expect accurate portrait AI.

A wide choice of artistic styles, with new themes added regularly. Good detail when the source photo is high quality. Successful results preserve facial recognizability.

A male face with long hair gets feminized. Group photos do not work. Over time styles begin to repeat and variety drops.

Fans of art avatars and social profiles who want an artistic look rather than a realistic photo.

Facial tracking is smooth and responsive. The AI companion receives positive reviews for creative dialogue. A one-time payment for pro access with no recurring fees is noted as a fair model.

After switching to a no-livestream mode, the live audience interaction that motivated many users is gone. Hand tracking is unstable. The app randomly signs users out during a session.

People who want to try mobile VTubing without investing in equipment. Works for recording short videos with an avatar, but not for those seeking a live audience.

AR filters run smoothly and look clean even with movement. A wide selection of beauty masks and effects. Lets you record video with filters without lag.

Visibly lightens and desaturates dark skin tones. Photo resolution is reduced without PRO access. The app has become noticeably laggy since AI features were added.

Users who want a camera with live AR masks and beauty filters for shooting video and selfies in real time.

A convenient collage tool with a large selection of layouts and fonts. Photo overlays and cutouts are used in a business context. Stickers and text work reliably.

The skin smoothing feature leaves scratches and artifacts. Reshaping facial structure gives unpredictable results. There is no meaningful AI processing: the app is primarily a collage editor.

People creating collages and designing posts for personal or business use, with no need for AI generation.

Manual sliders for individual features give real control over the result. The object removal tool works well. The basic features are enough for everyday editing without extra fees.

The app freezes on launch and requires several attempts before it opens. The background replacement tool only works for one face when multiple people are in the photo. Video features are unstable.

People who want to subtly correct a real photo without creating a synthetic avatar.

Skin retouching and hair color changes work quickly right during capture.

Smoothing is set to maximum on its own, and the face detector misses faces in groups.

Anyone who wants light selfie retouching for social media without deep editing.

Hair color changes in realistic tones look natural. Results with common haircuts on a clean face photo read convincingly.

Struggles with curly and afro textures. Dark skin produces unpredictable results. Long-term users consider the new AI interface worse than the previous neutral filters.

People who want to try a new color or cut before visiting a salon, with straight European hair types.

At its best, glossy quality resembling editorial photography, the 15-photo upload option for better likeness, and a simple intuitive interface.

After model updates, faces often bear no resemblance to the original. Extra hands and body distortions appear. Little content for men and styles are built around a female audience.

Women creating social media content who want glossy looks without a photo shoot, assuming the model quality does not degrade again.

The object-from-background cutout tool is one of the best on mobile. Layer blending and background replacement work cleanly. The image outpainting tool works convincingly.

AI headshots often miss the likeness: facial features, type, and expression all shift. A daily attempt limit sharply restricts usage. Long-time LightX users lost previously available editor features when the app was renamed.

Users who need a mobile editor with precise cutouts and blending. Less suitable for those seeking a realistic AI portrait with their own face.

With proper model training the face comes out recognizable and polished. Large style selection. Dark skin tones are rendered better than on competing apps.

Three arms, six fingers, disproportionate bodies: artifacts appear frequently. Generated bodies can look unrealistic in shape.

Users willing to invest time in training a model in exchange for polished high-quality looks for social media.

Single-face video swap quality at good lighting and a straight angle is described as the best among mobile alternatives. Recognition accuracy is high with clear source material.

After an update results became disturbing: faces distort beyond recognition. Multi-face swaps work unpredictably: faces get mixed up or incorrectly overlaid. Video processing is capped at one minute.

People looking for a tool to swap one face in a short video clip for entertainment or creative projects.

Fast generation with no long wait. Some headshots look professional and realistic. A wide variety of themed styles and backgrounds.

In a third of cases the face in the result does not resemble the person from the photo. The store page promises features like group photos that do not exist in the app. A large watermark remains on images.

People who want to experiment with themed photo looks and will not be bothered if a few shots miss the likeness.

Lip sync with voice is convincing. Several experienced business users report real time savings on content production. The interface is intuitive for beginners and allows a fast start with no technical knowledge.

After an update, avatar movements became more robotic and less cinematic than before. Generation limits hit unexpectedly quickly: a few videos in and the counter resets. The mobile version is significantly stripped down compared to the web version.

Business video creators, marketers, and small teams who need to produce presentations without live filming. Not for those seeking photorealistic results with their own face.

Sharpness and detail in old photos genuinely improve, and processing is fast.

The new algorithm changes facial features beyond recognition and produces a plastic look.

Anyone who needs to restore old family photos. Results are inconsistent.

Enhancing dark and blurry photos produces visible results without excessive smoothing. Portrait retouching preserves natural skin texture.

AI look generation is unstable: the result may bear no resemblance to the user. High load slows processing of high-quality images.

People who want to improve old or poorly shot photos, not those trying to create AI portraits.

Expressive 3D avatars with rich customization, the ability to stream, and to post videos.

Does not look like you: the cartoon stylization is far from a real portrait.

Teens who need a virtual identity for social interaction, not a realistic photo.

Enormous variety of effects and styles, fast processing, and the ability to make holiday cards with multiple people in a photo.

The face in the result often does not look like you. The AI pushes toward an averaged template. With darker skin the result sometimes changes ethnicity. Background and hand detail quality is low.

People who want to quickly make a creative collage or themed card without needing accurate portrait likeness.

Restoring and colorizing archival photos works convincingly. Removing unwanted background objects is effective. A wide set of collage tools and basic color correction.

AI portraits do not capture appearance: faces come out generic and far from the original. Skin retouching is too aggressive and naturalness is lost. Text-to-image generation frequently ignores prompt details.

People who want to restore old family photos or make a collage, but not for creating professional portraits.

Clarity mode delivers an impressive result on landscape shots with a single tap. Manual control over every shooting parameter in one interface without switching screens. Long-term users describe years of productive use.

The app does not open without an internet connection, making it useless when shooting in nature or while traveling. Several users report being unable to save photos to the gallery after updates. The editing extension from the native Photos app stopped working.

Mobile photographers who need manual shooting controls and RAW support on iPhone. Not suitable for those who shoot without a stable internet connection.

Simple templates produce a recognizable face and realistic animation without limb doubling. The result arrives quickly and genuinely makes people laugh.

Likeness is unstable: faces shift on complex poses. In pair videos one of the people may look completely different. A child's gender is sometimes misidentified.

People who want to share a fun themed video on social media without expecting photo-accurate likeness.

Accurate background removal for product photos, and batch processing saves time.

Virtual models do not resemble real people and are unsuitable for commercial use.

Marketplace sellers who need background removal, not portrait AI.

Cartoon and artistic filters work convincingly. The photo enhancement feature genuinely restores detail in blurry images.

Annual yearbook-style photos often output a different person. Curly hair is replaced with straight. Dark-skinned users have their skin lightened without reason.

A younger audience that wants quick artistic transformations and filters, not realistic portraits.

The object cutout tool is accurate and faster than competitors. A large font selection for design. Some AI backgrounds are generated convincingly.

Avatars poorly represent the user's appearance and repeat the same poses. AI text-to-image generation ignores prompt details. The app crashes mid-edit.

People making simple graphic collages who need cutout and overlay tools, not realistic portrait generation.

A variety of styles and character customization. Daily free attempts without spending money. An active community.

Male prompts almost always produce female images. The content filter blocks innocent words in prompts. Generation quality has noticeably declined with recent updates.

Anime fans and character creators for role-playing games who are comfortable with strict content moderation.

Quickly and amusingly animates any photo to a familiar tune. The result immediately gets laughs from viewers. Photos stay on the device and are not uploaded to servers.

Animation is limited to the mouth only: eyebrows, gaze, and head movements do not change. On some photos the face blurs and distorts. The result easily looks like a cheap edit under close inspection.

People who make memes and want to make an old photo 'sing' for a laugh in chats.

A rich selection of artistic styles with the option to create custom ones. An active community with contests. Daily free attempts allow trying the app without spending money.

The enhancement feature consistently lightens dark-skinned characters to white. Prompts are executed unreliably: specified elements disappear and unwanted ones appear. The content filter triggers on neutral words.

People creating stylized illustrations and concept art who are willing to spend time dialing in prompts.

Filters with simple movements and holiday scenes transfer facial features recognizably. The result draws laughs and joy, and is good enough for home viewing.

Faces often distort on complex poses. The image looks obviously generated. With a bad source photo the result looks aged, acne-ridden, or unrecognizable.

People who want a funny clip for family or social media without expecting accurate likeness.

Fast cropping for social media with the correct aspect ratios preserved.

The AI outpainter runs automatically and adds artifacts, with minimal user control.

Anyone who just needs to resize a photo for Instagram rather than portrait AI.

Artistic stylization quality, especially for anime and Renaissance portraits, is above average. Processing takes seconds when the servers are up.

Only one face per photo. Regularly changes skin color and eye color to lighter shades for people with a dark complexion. The server is unavailable for hours during peak times. Some popular styles were removed in recent updates.

People who want a one-off artistic avatar in a specific style, such as anime, Renaissance, or a Disney character, for a profile or as a gift.

A wide catalog of artistic themes: anime, fantasy, historical looks. The old photo restoration feature works for family archives.

Realism is sacrificed for stylization. Dark-skinned users are turned light-skinned. Artistic filters heavily alter appearance and no recognizable likeness remains.

Fans of artistic stylization and anime looks who care about style rather than portrait accuracy.

Previously generated 200 portraits with recognizable likeness and a variety of styles.

The new model does not preserve skin color, the mouth is always open, and style choices are limited.

Nobody in particular right now: the app has lost its niche between realism and art avatars.

When a generation succeeds, artistic quality is above average for the market. The enhance tool genuinely sharpens old photos. Supports a range of styles from realism to surrealism.

Does not follow precise prompt instructions, especially for hand details and backgrounds. A strict content filter blocks neutral requests. The interface has deteriorated significantly through a series of updates.

Designers and illustrators who need an unconventional stylized image rather than an accurate portrait of themselves.

Drawing and text tools work for creating simple clothing items for game avatars. Many fonts are available in the free tier.

Almost every template is locked behind PRO access. After an update, AI fully replaced manual creation and some users can no longer make clothing themselves. Publishing to Roblox is unstable.

Young Roblox players willing to pay for a subscription to create their own avatar clothing.

When face swap works, the face in the result is recognizable and close to the original. The algorithm consistently preserves features in good lighting.

Headshots almost never resemble the real person. The NSFW filter triggers on neutral and classical photos. Hairstyle choices are limited to one or two options in the base mode.

Hard to define: for most users the product does not do what it claims.

Cartoon stylization works quickly and produces a recognizable character when the source photo has a clear face in good lighting. The Disney filter delivers quality results with proper exposure.

The algorithm consistently makes dark-skinned people lighter and changes hair texture to afro regardless of the original. Refuses to process bald heads, hats, or partially obscured faces. Facial expression and details (eyes, nose) are often lost in simplification.

People who want a one-time fun avatar in a cartoon style for a messenger, with no expectation of portrait accuracy.

Lets you quickly look 'retouched' in video and photos for social media. An intensity slider gives control over the degree of processing. Works in real time on video.

Lightens dark skin and smooths facial features beyond recognition. Eyelash artifacts are visible on every filter without close inspection. Reduces the resolution of the final image.

Social media content creators who need a quick beauty filter on video without aiming for realism.

Some users rate face alignment accuracy highly in the basic mode and enjoy the variety of templates: one reviewer wrote that in 30 seconds the face matched perfectly. The gender swap mode is seen as an entertaining feature.

Results for dark skin are catastrophic: the app either lightens the face beyond recognition or refuses to process the photo at all. Faces blur and distort on overlay, glasses disappear, and images do not align when using multiple photos.

Light-skinned users who want a quick themed look for entertainment.

A large variety of artistic modes including animation and background replacement. Several users praise the interface simplicity and speed on good days.

The result rarely resembles a specific person. Skin becomes too light and facial features are generalized. One reviewer: 'Endless style choices, but every time the output is a stranger'. Racial skin shift is mentioned by several users who are not white.

People looking for artistic photo stylization for the effect, not for a recognizable portrait.

A simple interface and basic skin smoothing work for uncomplicated portraits and light retouching.

Body shape tools produce visible outline distortions. Some effects do not apply when tapped. The app crashes when opening photos from recent iOS versions.

For light skin retouching on simple portraits, if you can sort out the payment flow.

Film simulations produce a convincing analog look. Real-time skin correction on video looks natural. The video quality enhancement tool noticeably outperforms most competitors.

The algorithm automatically removes moles and marks with no way to keep them. Glow Up sometimes makes faces comically smooth. Daily limits appear without warning even for VIP users.

Bloggers and short video creators who need fast video retouching while preserving a natural look, not a radical transformation.

In the best cases, improving the resolution of old family photos yields a result good enough to print. Colorizing black-and-white images works convincingly for neutral landscapes.

Results depend on the photo, and faces often look like strangers. Frequent server outages interrupt work. The aging filter and 'baby filter' produce characters that bear no resemblance to the original.

People with a specific old blurry photo they want to improve for printing, who are willing to try several times to get a usable result.

Applying makeup over a photo through a live camera works convincingly. A wide set of tools for skin tone and highlight. Correction results at moderate settings look natural.

AI removes moles and marks without user permission and with no way to preserve them. Face correction quickly becomes unnaturally smooth. Collage tools run slowly even on powerful devices.

Users who want quick cosmetic retouching for beauty content, if they are willing to monitor subscription terms carefully.

A rich library of artistic styles, fast rendering. Some results look stylish. Repeat generation is available with one tap.

People with dark skin, Asian, and Filipino features consistently come out white in results. The watermark remains on exports. Video rendering takes up to 20 minutes for a 10-second clip.

Experimenters with artistic filters who are comfortable with approximate likeness and have no need for clean exports.

In rare successful generations the face looks convincing on a dance video. Processing speed is acceptable.

The face in the result often does not resemble the uploaded photo. Several users describe the output as a completely different person. The AI removed clothing from a child when generating a mermaid look from a photo with a covered torso.

An extremely narrow audience: people who specifically want an entertainment video effect with a dancing avatar and can tolerate low accuracy.

Sometimes produces realistic LinkedIn portraits. Good variety of styles and themes. Hands without obvious anomalies, which is rare in this category.

Regularly changes the user's race, gender, or hair color. Neck and head proportions distort. Results are unpredictable: one good attempt is followed by three bad ones.

People who want a casual fun portrait and are comfortable with lottery-style results.

The 'live photo' concept with controlled movement was strong, and results for simple nature or object animations sometimes come out atmospheric.

Video generation takes 10-24 hours and frequently outputs a stranger's face or a black screen. Men are turned into women in the output. Results are described as 'amateur' even in successful generations.

Practically no one in the current state: the product does not fulfill its basic promise.

Retouching and object cutout tools are accurate enough for everyday tasks. A wide feature set in one place.

AI avatars do not resemble the original. Portrait effects produce a plastic, puppet-like look. Anime and cartoon styles distort facial features and lighten dark-skinned users.

People looking for an editor with basic retouching and cropping, not accurate AI portraits.

At its best, detailed images that users compared to Midjourney. Several artists praise it as a tool for quick concept visualization.

The generator ignores specific instructions about poses, clothing, and scene composition. Results almost always show faces even when a landscape is requested. Generative fill quality has degraded to a cartoon level according to several independent reviews.

Artists who need quick rough sketches, using the web version rather than the mobile app.

In lucky cases it generates convincing realistic looks for product photos. A wide range of artistic styles, some of which produce interesting results.

The output regularly does not resemble the person in the photo. Prompts are ignored: the app does what it wants. Every update removes styles and features (Schnell, SDXL, graphic novel, AI influence control) that users paid for.

Authors and illustrators who need generalized artistic character looks, not an accurate portrait of themselves.

Ready-made story templates speed up production of standard content. For experienced users the interface is flexible enough to produce acceptable results.

AI generation produces anatomical errors: extra arms and distorted faces. Prompts are only partially followed or ignored entirely. The gap between what is shown in the app store and the actual product is critically large.

SMM specialists who need a fast flow of templated story content, with no expectations from AI quality.

Some users are happy with the collection of high-resolution wallpapers for iPhone. The selection of background images is diverse.

The AI generator ignores or misinterprets prompts. Results featuring people contain distorted facial features. The attempt limit is too small to assess quality before the first payment.

Anyone looking for phone wallpapers, if this is treated as a wallpaper catalog rather than an AI tool.

The variety of themed packs attracts people who want to experiment with different looks. A handful of users received results they considered suitable for a professional profile.

The photo does not resemble the person: the app changes ethnic features, hair color, age, and body size without being asked. One user received a photo in club attire on horseback instead of a business shot for a realtor profile. Real facial features are barely detectable in the output.

Practically no one with consistent expectations of matching the original.

Fast rough renders for storyboarding and concepts. A relatively wide range of styles for text-to-visual experiments.

Images are blurry and incoherent with detailed prompts. Characters change gender, ethnicity, or gain extra items. Iterative editing does not work: a repeated prompt is ignored.

Content creators who need rough conceptual sketches, not precise portraits.

Isolated cases of restoring heavily damaged or very old photos (nearly 100 years old) are described as impressive. The pixel-sharpening feature on pixelated images sometimes works well.

Faces after processing become cartoonish or belong to different people. A sharpei dog photo became a different breed with different coloring. Portrait retouching ignores the person's original appearance.

Practically no one in the current state.

Individual users successfully restored old black-and-white family photos and colorized them. Makeup templates on some photos produce a clean and natural result.

After processing, the face often does not resemble the person in the source photo. Strangers appear in the frame and body parts are cut off. Attempts are consumed even on failed generations with no option to retry.

May work for restoring old family photos. Not suitable for people who want to try a new hairstyle or get a portrait with recognizable features.

For light-skinned users, facial features are transferred recognizably and the classic portrait painting style is convincing.

Systematically whitens dark skin and changes nose shape, eye shape, and iris color. For most users who are not of Asian or European background, the result is unrecognizable.

Only for light-skinned users with European features who are interested in a classical painting style.

In the best cases the video looks cinematic with natural movement. The result impresses those who have figured out the right prompts.

The final video often ignores prompt details. The content filter triggers on neutral scenes. The generation does not match what was written in the request.

Content creators with a budget who are willing to experiment and accept lost credits for the occasional good clip.

Several users received realistic paired portraits during the holiday period. Generation speed is praised as higher than competitors.

Extra legs and three fingers are a recurring defect. The app ignores glasses despite photos showing glasses. The output does not resemble the person, with faces looking glossy and generic. Racial shift is mentioned by several users.

People who are comfortable with experimental quality and do not expect portrait accuracy.

The concept of animating old photos is appealing as an idea. A few users noted the first or second result was acceptable in quality.

Faces in animated clips look unnatural and the smiles are eerie. The app changes facial features beyond recognition, especially when shifting ethnic appearance. Generation can take 4-7 hours with no guarantee of a result.

No one until speed and appearance accuracy problems are fixed.

For non-portrait subjects such as backgrounds, objects, and animals, enhancement sometimes works acceptably.

Faces are systematically distorted or replaced with strangers. Video enhancement adds cartoon animation instead of clarity. Old photo restoration disfigures faces.

No one who expects faces to remain recognizable in processed images.

Some users got acceptable results for short dance clips where the focus is not on a specific face. The interface is relatively straightforward.

The face from the uploaded photo is almost never preserved in the final video, and sometimes gender and overall appearance change completely. Movements look unnatural. Use of strangers' faces in promo materials is documented in reviews as a violation.

Practically no one who expects likeness to the real person in the photo.

Individual users found a good shot among a pack of four, especially in early versions when the algorithm was more stable.

The result does not resemble the person in the photo. Hands and fingers are distorted, extra legs and random strangers appear in shots. Skin looks plastic and head-to-body proportions break. One review: '99% of photos look like a completely different person'.

Effectively no one at the current quality level.

The idea of themed photo sessions without actual shooting is appealing. Occasionally the first generation is acceptable with a very good source photo in bright lighting.

Faces are distorted or unrecognizable in almost every set. The body in the generation does not match the user's real figure. Quality is significantly below competitors at a comparable price.

Practically no one at this stage: the product does not reach its stated quality.

In rare cases the basic photo animation works correctly and is touching, especially for photos of elderly relatives.

AI avatars do not preserve likeness: eye color, hair color, and skin tone do not match. Processing hangs during generation and does not complete for hours. Pet video content barely works.

Practically no one: the ratio of result quality to stated functionality is unacceptable.

Early versions running Stable Diffusion delivered acceptable results for some creative tasks, according to longtime users.

Images with people consistently contain anatomical defects: extra limbs, distorted faces, and prompt mismatches. One user submitted 60 requests and got zero usable results. Prompts are ignored or interpreted arbitrarily.

Effectively no one in the current state: output is too unpredictable.

In the best cases, artistic styles such as painting and anime look striking regardless of likeness accuracy.

Faces do not resemble the uploaded photos. Gender is misidentified. Repeat packs contain duplicates of already received images. Removing account access through the app menu is not possible.

Hard to recommend: too many complaints about basic non-resemblance and payment issues.

The idea of interacting with a departed loved one is emotionally significant. When a video does come through, one user called it 'the closest to reality'.

Faces in the output do not resemble the uploaded photos. Hands are distorted and bodies are out of proportion. Generation takes 10 to 40 minutes and frequently ends in an error.

Theoretically for people who want to animate archival photos of relatives, but in practice the result is too unreliable.

Individual users noted acceptable Halloween looks on simple templates with a good source photo.

The result is rarely recognizable as the user. The face is placed on an inappropriate body. Hair and distinctive features are ignored. Effects look unappealing.

Practically no one: quality does not justify even free use according to real buyers.

Some users got acceptable results for short dance clips with no emphasis on a specific face.

Faces do not resemble the uploaded photo, and sometimes gender and overall appearance change completely. Movements look unnatural. Use of strangers' faces in promo materials is documented in reviews as a violation.

Hard to recommend to anyone given the current quality and reputation.

A variety of styles and themes, vibrant visual aesthetics in results. When they come out, some people find the images beautiful in their own right.

The face and appearance of the person in the source photo are not transferred to the output. Output videos look staged and unrealistic. Anatomical deformations such as extra limbs appear. No control over style or background.

Not recommended for anyone as a face-recognition tool.

Frames and background effects without face placement sometimes look acceptable.

Faces are completely replaced with strangers bearing no resemblance to the original. Movements are robotic and unnatural. Two-person scenes from store materials output a static image instead of animated interaction.

No one who expects even a loose resemblance to the original photo.

The idea of seeing yourself at different ages or in different styles generates interest: several users found the concept enjoyable.

Hairstyles in unrealistic colors, aged results with terrible teeth, a cartoon look that does not resemble you, and dark skin whitened beyond recognition. Photos of plants and personal photos are rejected as 'illegal content'.

Practically no one in the current state.

Rare successful results on front-facing portrait photos of good quality look lively and convincing.

Faces do not resemble the original or are completely replaced by a different person. Animation often reduces to a head nod or zoom with no real movement. The racial appearance of characters in some photos changed during processing.

Only for those willing to spend money on the chance of a rare successful result from a perfect photo.

The caricature style of images appeals to some users. The idea of trying on different professions in a portrait is perceived as a fun concept.

The result does not resemble the person in the photo, and this is a systemic problem not an exception. Dark skin is rendered with noticeable distortions. Bodies are out of proportion and hands are enlarged.

A fun generator of 'yourself in a different look' with no expectation of portrait accuracy.

In isolated cases with an optimal photo set, facial likeness is good enough for use as an avatar in a work profile.

The result almost never resembles the user. Cartoonishness, finger and body distortion, phantom hair on bald people, and whitening of non-white features are consistent problems. There is no way to see the result before paying.

No one: quality does not justify a paid subscription according to the overwhelming majority of buyers.

Rare successful results look convincing and realistic.

The face and appearance from uploaded photos are not preserved: different hair color, different body type, different racial features.

Only suitable for people who are fine with a good-looking avatar that does not have to resemble them.

The concept of a quick professional headshot without a photo studio is appealing. In rare cases users noted '2-3 decent shots out of 20'.

Portrait likeness is absent: a different person in the output is the rule, not the exception. Different head and body skin tones. Anatomical artifacts (extra limbs, distorted features). Poor performance with dark skin.

Not recommended for anyone in the current state.

Occasionally stylized frames came out impressively and those who landed a good generation noted a vivid visual style. The concept of age progression and photo animation is appealing in itself.

The result almost never resembles the person in the uploaded photo: different ethnic appearance, different gender, different age. Animation reduces to a light zoom with no real movement. Processing hangs for hours and frequently ends in an error.

Practically no one who wants a personal result. A random entertainment effect with no connection to a specific person.

The idea of animating photos with text control is appealing. Individual users received simple looped animations for background reels.

The result does not match the uploaded photo: a different person, a different object, a different style. Video frequently gets stuck on generation and does not complete. Quality resembles a low-quality animated drawing, not realistic video.

Practically no one for getting a predictable result. Not recommended.

Individual users who managed to receive a finished video note pleasant stylization. The concept of a trending video from a single photo is appealing as an idea.

Generated faces do not match the source photo: different type, different features. Generation frequently hangs and does not complete. A weekly limit of a few videos even on an annual plan makes the app practically unusable for regular use.

No one who expects to recognize themselves in the result. Perhaps for a one-time experiment with the understanding that the output will be arbitrary.