I Love the Smiths Video Generator: Put Two People in the Elevator Scene

Which cases are returned

• Returned automatically: any run that errors out before a video exists.
• Not returned: a finished clip in which a face changes from its photo, the doors look wrong or a third person appears. It counts as an outcome, not a broken run.
• Trying again: a different photo or option takes 1 credit, and a 12-second try is one of the lower-cost runs here.
• An unopened pack: may go back for a refund within 7 days under the refund policy.

What to look at in the preview first

Make sure each face is recognisable, there are only two people and nothing is written anywhere. The fix guides cover the common misses: the faces change in the elevator (guide), the headphones or the elevator look odd (guide), and a third person or some text appears (guide). Previews per pack: 1 ($6.99), 3 ($10.99) and 5 ($15.99).

To request a refund for an unused pack inside 7 days, write to support@mimiczip.com.

Quick fixes

• Pick a photo where the eyes are open and the face is fully visible.
• Use even light, without a strong shadow across the face.
• Start with the metal elevator and your own clothes before changing both.
• Keep hair, hats and scarves off the face in the photo.

Why it happens

Two faces close together, a turning head and mirror-like walls change the picture a lot. Sharp photos give the model clean faces to keep.

Quick fixes

• Use a whole-body photo with a clear outline.
• Start with the metal elevator and the crush mood.
• Pick the office elevator next: flat walls are easier than mirror-like steel.
• Run again rather than changing several options at once.

Why it happens

The model has to draw headphones, two faces, reflections and moving doors, and it redraws all of them every frame. A second run often draws them cleaner.

Quick fixes

• Crop out other people and pets.
• Avoid shirts with printed words or logos.
• Avoid photos with a poster, a sign or a screen behind you.
• Use a plain background.

Why it happens

The model sometimes turns a second face in a photo into a third person, and printed words on a shirt can be redrawn as a sign. A clean photo of one person avoids both.

Editing it yourself or generating it

• Where an editor helps: cutting, trimming and captioning footage you own, or dropping a face into a template. It cannot put you into an elevator scene starting from stills.
• What MimicZip does: it creates a two-person elevator scene with headphones, two lines and an opening door and makes your photos the stars.
• What goes in: one photo of each of the two adults.
• What you get back: a 12-second 720p clip that follows the whole 14-second TikTok by @battos13 from 5 November 2026, shortened to 12 seconds, with no text on screen.
• Price: CapCut offers free and paid features. Here one credit is one full video without a watermark (a two-person elevator scene with headphones, two lines and an opening door), and every pack includes watch-only previews.

Choosing between them

Edit by hand to cut a specific clip. Generate with MimicZip for a two-person elevator scene with headphones, two lines and an opening door made around your photo. You can also generate here and add text in CapCut.

Where a prompt struggles with this scene

• What you type: a general tool wants a prompt and settings. MimicZip has no prompt box; you add a photo and choose the options.
• Your look: a general tool often changes a face when the head turns or a reflection moves; our scene is built to hold yours.
• The beats: the listening, the first line, the second line and the doors need the right order, and we fix the order for you.
• Nothing borrowed: the prompt never names a song or an account, which many tools filter anyway.
• Cost: general tools run on their own plans. MimicZip has packs from $6.99, 1 credit per clip and previews in each pack.

When a general tool is better

Use a general tool for your own story or full camera control. Use MimicZip when you want exactly this clip from your photo, fast: a two-person elevator scene with headphones, two lines and an opening door. To write the prompt yourself, see Kling, Veo and Seedance.

Copy this prompt

Image-to-video, 16:9, about 12 seconds.
Reference image 1 is the listener and reference image 2 is the talker; both are adults. Keep each one's face, hair, skin tone and body shape exactly in every second. Both wear their own everyday clothes from the photos. The mood between them is a shy, warm crush: the talker is a little nervous and smiling, the listener is surprised and charmed.
Setting: inside an old brushed-metal elevator with a dark panel of buttons, mirror-like steel walls and soft warm ceiling lights. One continuous take, a fixed medium shot from the elevator wall, both in view side by side, no zoom, no cuts.
0-3 s: the two stand side by side, the listener wearing headphones and looking ahead; the talker glances sideways at the headphones.
3-6 s: the talker turns to the listener and says something, with a warm, shy look; the listener slides one side of the headphones off and turns a little.
6-9 s: the talker steps a little closer and says one more short line with a smile; the listener smiles back.
9-12 s: the elevator doors slide open, the talker steps out and glances back once, and the listener watches with a happy smile.
The mouths move, but nothing is heard. No text, numbers or logos anywhere. Audio: elevator hum and one door chime only, no music, no vocals. Photorealistic, natural colour, no subtitles, no logos, no watermark.

How to use it

• Attach two photos, one of each adult as the reference image(s).
• Add your sound in TikTok afterwards. A text prompt cannot place each beat exactly, so expect to shift the sound.
• Leave names, sound titles and brands out. Tools can filter them, so the prompt describes the action instead.
• Tools change often, so check the length and aspect ratio your tool supports and adjust the first line.

Copy this prompt

Create a 12-second cinematic video from the reference image, 16:9.
Reference image 1 is the listener and reference image 2 is the talker; both are adults. Keep each one's face, hair, skin tone and body shape exactly in every second. Both wear their own everyday clothes from the photos. The mood between them is a shy, warm crush: the talker is a little nervous and smiling, the listener is surprised and charmed.
Setting: inside an old brushed-metal elevator with a dark panel of buttons, mirror-like steel walls and soft warm ceiling lights. One continuous take, a fixed medium shot from the elevator wall, both in view side by side, no zoom, no cuts.
0-3 s: the two stand side by side, the listener wearing headphones and looking ahead; the talker glances sideways at the headphones.
3-6 s: the talker turns to the listener and says something, with a warm, shy look; the listener slides one side of the headphones off and turns a little.
6-9 s: the talker steps a little closer and says one more short line with a smile; the listener smiles back.
9-12 s: the elevator doors slide open, the talker steps out and glances back once, and the listener watches with a happy smile.
The mouths move, but nothing is heard. No text, numbers or logos anywhere. Audio: elevator hum and one door chime only, no music, no vocals. Photorealistic, natural colour, no subtitles, no logos, no watermark.

How to use it

• Attach two photos, one of each adult as the reference image(s).
• Add your sound in TikTok afterwards. A text prompt cannot place each beat exactly, so expect to shift the sound.
• Leave names, sound titles and brands out. Tools can filter them, so the prompt describes the action instead.
• Tools change often, so check the length and aspect ratio your tool supports and adjust the first line.

Copy this prompt

Generate a 12-second 16:9 video using the reference image as the subject.
Reference image 1 is the listener and reference image 2 is the talker; both are adults. Keep each one's face, hair, skin tone and body shape exactly in every second. Both wear their own everyday clothes from the photos. The mood between them is a shy, warm crush: the talker is a little nervous and smiling, the listener is surprised and charmed.
Setting: inside an old brushed-metal elevator with a dark panel of buttons, mirror-like steel walls and soft warm ceiling lights. One continuous take, a fixed medium shot from the elevator wall, both in view side by side, no zoom, no cuts.
0-3 s: the two stand side by side, the listener wearing headphones and looking ahead; the talker glances sideways at the headphones.
3-6 s: the talker turns to the listener and says something, with a warm, shy look; the listener slides one side of the headphones off and turns a little.
6-9 s: the talker steps a little closer and says one more short line with a smile; the listener smiles back.
9-12 s: the elevator doors slide open, the talker steps out and glances back once, and the listener watches with a happy smile.
The mouths move, but nothing is heard. No text, numbers or logos anywhere. Audio: elevator hum and one door chime only, no music, no vocals. Photorealistic, natural colour, no subtitles, no logos, no watermark.

How to use it

• Attach two photos, one of each adult as the reference image(s).
• Add your sound in TikTok afterwards. A text prompt cannot place each beat exactly, so expect to shift the sound.
• Leave names, sound titles and brands out. Tools can filter them, so the prompt describes the action instead.
• Tools change often, so check the length and aspect ratio your tool supports and adjust the first line.