Symmetry Filter: Explore a mirrored-face look as a video or a still.
Use a front-facing selfie for the symmetry-style edit, then choose a moving comparison or a single generated image. Lighting and camera angle strongly influence how the transformation reads.
The symmetry trend asks a simple visual question: how different would your face look if each half became the whole? One version mirrors the left half; another mirrors the right. Comparing those two constructed faces with the original can exaggerate tiny differences in eyebrows, hair, lighting and expression, which is why the result often feels like a pair of unfamiliar twins.
Smew makes the idea from one straight-on selfie. Choose a ten-second video that moves between the looks, or a single vertical photo strip with the two interpretations and the original. This is an AI visual effect, not a medical measurement, beauty score or exact pixel-mirroring utility. The distinction matters if your goal is a precise symmetry analysis rather than a social post.
Also called Asymmetry vs Symmetry, symmetrical face, mirror-face effect
Example
See the Symmetry Filter reference
Credited examples explain the trend’s staging. They are visual references, not Smew customer results; your generated details and movement can differ.
Trend reference
A reference for the scene and timing. This sample is not a Smew customer generation. Compare the framing and sequence rather than expecting a frame-for-frame duplicate.
Start with the right version and clear photos; the template provides the scene instructions.
Choose the scene and cast
Open Create and choose the version that matches the result you want. Check the photo-slot labels before uploading: one identity belongs in each slot. A reference sample shows the idea; the choices on Create define which people or pets Smew can actually place in that scene.
Add photos and review the output
Upload the requested photos, wait for processing and confirm that the preview shows the correct subject. Compare the selected output, model and credit total. You do not need to record the action or paste a prompt: the template supplies the setting and movement. Submit only when those details match your choice.
Follow the task, then review the finished file
After submission, follow the task on Dashboard while it generates; you can browse templates and see your other tasks there. Once it finishes, open the saved result in Library and inspect the whole clip before downloading. The generated interpretation can vary, especially during fast turns, occlusion and close contact.
Which photo gives this template a fair start?
Take the selfie with the camera at eye level and your head upright. Both eyes and cheeks should be equally visible; a turn toward one shoulder makes one side appear wider before the effect even starts. Keep the face large and sharp, with no hand, sunglasses or hair covering a cheek.
Use soft front light, such as facing a window. A strong lamp on one side can make one mirrored interpretation much darker than the other. A neutral expression or gentle smile is easier to compare than an open mouth. You do not need to remove your natural hair part or freckles; those differences are part of what viewers notice. Use your own adult photo or that of an adult who agreed, rather than a child or an unconsenting public figure.
The background of Symmetry Filter
Face-flipping and face-symmetry posts have appeared on social platforms for years. The older Inverted-filter discussion concerned seeing an entire face flipped; the later symmetry format instead builds a whole face from one side at a time. Those are related comparisons, but they are not the same operation.
The October 2026 references include tutorials for a CapCut symmetrical-face template and posts comparing asymmetry with symmetry. Their captions explain why users search for a mask filter, mirror face or four-picture template. Smew offers an AI interpretation of the visual concept, using one photo rather than requiring four uploads. The linked examples document how the topic was presented; they do not establish an objective standard for a face or a scientific ranking of appearance.
The older Inverted-filter discussion concerns flipping a whole face, a useful distinction from half-face mirroring. Grazia
A CapCut tutorial shows the symmetrical-face editing template associated with this wave. TikTok
A Spanish-language tutorial presents the same asymmetry-versus-symmetry idea to another audience. TikTok
A compilation collects symmetry comparisons; its role here is visual context rather than a popularity claim. TikTok
Copy and paste
The scene instructions are already included
The built-in recipe sets the setting, roles and sequence. Your photos provide the subjects; you can read the recipe below without needing to paste or rewrite it.
Video
1 photo: your selfie, face straight to the cameraSeedance 2.5 Video Edit7 s · 720p · 9:16126 credits
Video prompt
Edit [Video 1]. Replace only the explicitly mapped source subjects in place:
[Image 1] replaces each source foreground face to which the symmetry effect is applied; photo slot: Your selfie, face straight to the camera, in every shot where that subject appears.
[Image 1]: Use the uploaded subject's appearance as the identity base while retaining the original source effect and its progression. Preserve source costume reveals, hair/makeup changes, morph stages, deliberate stylization and occlusion; do not expose a face through a mask or force the untransformed appearance over the final effect.
Selected variant: Video / main. Apply only the requested cast, wardrobe, setting or effect changes for this variant; preserve the source action and shot timing instead of inventing a new scene sequence.
Use the complete preview as the source for the Symmetry Filter / Video variant. Preserve its actual action, effect progression, camera moves, shot sequence, physical interactions and mouth/jaw timing. Photo-role labels determine the intended identity assignment; never assign uploaded faces to unrelated background people or props.
Track each source subject through front, profile, back and occluded views. Keep all non-target subjects unchanged. Never duplicate an uploaded identity, merge subjects, swap their roles or transfer an identity to a prop or background subject.
Preserve the complete source video from its first frame to its last: source duration, frame order, every shot boundary, playback speed, camera movement, framing, lighting and original performance; keep background/props unless the selected variant explicitly changes them. Preserve each subject's original head, gaze, body, hand/paw gestures, expression, mouth and jaw opening/closing and their exact source timing. Do not create a new dance, scene, shot list, dialogue or lyrics. Do not add, remove, reorder, shorten, extend or retime shots or actions.
Produce silent video. The matching unmodified original soundtrack is restored separately after this edit. Do not generate speech, singing or music. Do not add new captions, subtitles, logos or a watermark.
1 photo: your selfie, face straight to the cameraNano Banana 2.1Photo · 2K · 9:163 credits
Photo prompt
Produce one 9:16 image. Use only the adult identity from [Image 1], without changing or idealizing facial structure. Base the personal likeness on the supplied photographs: retain face shape, complexion, age cues, eyes, hair and distinguishing details while applying the specified treatment. Do not substitute a recognizable public figure or invent a different identity.
Build one vertical image as three equal horizontal strips separated by narrow white lines. Keep the same straight-on face scale, hair-to-collarbone crop and lighting in all strips. The top strip is made only from the picture-left half of the face, mirrored across the nose and chin. The middle strip preserves the real face. The lower strip uses only the picture-right half, reflected toward the other side. Mirror the corresponding background and shoulders in the outer strips. Keep natural texture and asymmetry in the ordinary view; do not create a seam, score, measurement graphic or a different person.
Keep the composition readable, with correct separation between subjects, consistent lighting and plausible anatomy. Preserve intentional stylization without accidental duplicated faces, disconnected hands or merged bodies. If an input is an animal, preserve its own species, facial features and fur instead of adding human facial skin. No additional captions, labels, logos, signatures, watermark or interface elements.
You pay in credits. The Generate button shows this price before you spend anything. These are the options on this page.
Option
Model
What you get
Credits
Video
Seedance 2.5 Video Edit
7 s · 720p · 9:16
126
Photo
Nano Banana 2.1
Photo · 2K · 9:16
3
One-time packs and plans are on the pricing page. A failed video gives its credits back.
Video comparison or a single photo strip
The video starts with the source identity, shows the left-based and right-based interpretations, then returns to the familiar face. The point is the contrast, so the transitions should leave enough time to recognize each version. The photo choice places the same idea into one vertical composition: a mirrored interpretation, the original, and the other mirrored interpretation. It is useful when you want viewers to compare without pausing a clip.
An AI model draws the result rather than mechanically copying a column of pixels. Hair strands, moles and fine edges may move, and the intended mirror line can be imperfect. If you need a mathematically exact flip for design or measurement, use an image-editing tool. Choose this template for a stylized social comparison, with the actual outputs below.
What asymmetry versus symmetry is showing
Most ordinary portraits contain visible differences between the two sides: a raised eyebrow, an uneven smile, a hair part or a shadow. Mirroring one side repeats those details and can make them feel more dramatic. It does not follow that the original face is faulty, or that one generated version is a more correct appearance.
The effect provides no numerical score and should not be used to diagnose a condition, judge attractiveness or decide which side of a face needs fixing. A caption can keep the tone curious: ask which version looks least familiar rather than inviting a beauty ranking. Because the model also interprets the image, differences between the outputs are not a reliable measure of your natural facial asymmetry.
The current Smew options are listed below. The photo count belongs to the selected version: an extra person in a single upload is not a substitute for an additional photo slot. The model name describes how the output is made; it does not promise identical faces or motion on every run. Review the selected option and credit total on Create before submitting.
Choice
Photos
Model
Output
Credits
Video
1
Seedance 2.5
10 seconds, 9:16, 720p
126
Photo
1
Nano Banana 2.1
9:16, 2K
3
Sound varies by choice. Video uses sound generated with the scene, rather than a promised copy of a social-platform recording. Photo uses a still image, without an audio track. A reference clip’s audio is not evidence that every option includes it. Music rights remain with the rights holders.
An AI interpretation is different from a mirror mask
A conventional editor can cut an image at a center line and flip the chosen half directly. That operation preserves the original pixels, even if the seam is visible. Smew’s template instead describes the two mirrored looks to an image or video model. The model has to redraw the face and background, so the result can be visually close without being exact.
You do not need to locate a creator’s prompt or reconstruct a closed editing template. The scene instructions are already included. The important input decision is a level, well-lit face. A better prompt cannot recover an eye hidden by hair or undo severe side lighting; changing the photo is more useful than repeatedly asking the model to infer missing information.
Fix the photo before repeating the comparison
A wide and narrow pair usually begins with a turned head. Unequal brightness often begins with side lighting. If either appears, take a new front-facing selfie rather than treating the output as a discovery about your face. The input itself can create most of the dramatic contrast.
A central seam, altered mole or distorted hairline may be a redraw artifact. Inspect the original and the generated versions together. If the faces barely differ, that is not a failed beauty test; the source may simply have few visible differences, or the model may have softened them. Avoid spending repeatedly in pursuit of a supposedly correct score, because this page does not calculate one.
FAQ
Symmetry Filter — questions before you create
What is the symmetry filter?+
It creates two face interpretations, one based on mirroring each side of your selfie, and compares them with the original. The trend is often captioned Asymmetry versus Symmetry. Smew offers a generated video or a still photo strip from one adult portrait, rather than a live camera lens.
What does "Asymmetry vs Symmetry" mean in the trend?+
The original portrait contains natural differences between its two sides. Each mirrored interpretation repeats one side across the center, making it look more symmetrical. The contrast is a visual game, not a judgment that one face is better. AI redraws can also introduce differences that were not in the source.
How do I get the symmetry filter on TikTok? Is it a CapCut template?+
It is a separate Smew interpretation of that trend. The referenced tutorials show a CapCut symmetrical-face template, sometimes described as a four-picture mask. Here you upload one selfie and select Video or Photo. You do not need the editor template link or four input photos.
Is there a symmetry filter app or a symmetry filter online?+
Yes. Create runs in your browser. Upload a clear straight-on selfie, review the selected output and credit amount, then submit. The finished image or video is saved for review. It is a generated file rather than a camera filter you watch in real time.
Is this a face symmetry test?+
No. It does not measure landmarks, calculate a ratio or provide a medical or beauty score. It generates pictures that illustrate the mirrored-face idea. Camera position, light and the model’s redraw affect the result, so it cannot reliably measure facial symmetry or diagnose anything.
Is there a symmetry filter on Instagram or Snapchat? What is the filter's name?+
Names vary by app and creator. The source wave here is associated with symmetrical-face and asymmetry-versus-symmetry captions, not a verified universal Instagram or Snapchat lens name. Smew produces an independent file that can be shared in compatible posts; it does not install a lens in those apps.
What is the inverted symmetry filter?+
Inverting flips the whole image horizontally. Mirroring half a face duplicates one side to construct a new whole face. This template is about the second idea. The two effects can both feel unfamiliar, but they show different things and should not be used interchangeably as evidence of how others see you.
Do I get a perfectly symmetrical face?+
That is the intended visual direction, but the model redraws the face rather than copying pixels exactly. Small details or the center seam can be imperfect. Use a dedicated editing tool if exact mirroring is necessary; choose this template for an AI interpretation of the social comparison.
Is my face symmetrical? Can the filter tell me?+
It can show an entertaining comparison, but it cannot give a dependable measurement. A tilted head, unequal light or an AI artifact can exaggerate differences. Treat the outputs as generated images and keep the original in view, rather than drawing conclusions about appearance or health from them.
Can I use an asymmetrical face filter instead?+
The original identity is part of the intended comparison: the video returns to it, and the photo strip includes it between the mirrored interpretations. You do not need a separate asymmetry option. Check the completed file to ensure the model retained the intended comparison rather than smoothing every face into the same look.
Is it the same as a mirror face filter?+
It belongs to that visual family, because each interpretation uses a side of the face as its basis. The distinction is the AI redraw. A conventional mirror tool literally flips pixels; this template generates a new image or sequence from instructions and the selfie, so fine details can vary.
Who can use it?+
Use your own adult selfie or an adult who has agreed to this use. Keep one clearly visible face in the picture. Do not use the trend to judge someone who did not consent, and do not treat a child or celebrity example found elsewhere as a supported input for this template.
Does it cost credits?+
Video currently uses Seedance 2.5 for a ten-second, 9:16, 720p result at 126 credits, with generated scene audio rather than the reference recording. Photo uses Nano Banana 2.1 for a 2K vertical image at 3 credits. The selected total is displayed on Create before submission.
Is my photo private?+
Check Smew’s privacy policy before submitting an image and use one you have permission to upload. The public examples on this page explain the trend; they are separate from your photo. Another site’s statements about storage duration, deletion or sharing are not automatically Smew’s policy.
