I'm Sorry I Lied dream vs truth: a hard cut

How the dream-versus-truth cut is built

Your photo supplies the identity reference. The opening and final looks are generated from the selected template. A confession template inserts text over a blurred background; a progressive template generates a distinct middle image. Identity and scene details can vary.

Photo choices that make the contrast clear

Show the costume. The homemade outfit is the joke, so pick a photo where it is visible, head to toe if you have it.

Keep the face readable. Even light and an unobstructed face give the AI more to hold onto. Avoid sunglasses, heavy filters and hair across the eyes.

One subject. Use a single clear child. Confirm the person in the photo is you or your own child and that you have permission to share.

What to check when the reveal looks wrong

The face changed. Compare the dream slide with your source photo. A sharp, well-lit face helps, but identity can drift; try another photo.

The costume is off. The AI redraws the costume from a written description, so details vary. Pick the costume again or adjust the photo.

The slides look identical. The dream and truth should be clearly different. If they look the same, the costume may not have been applied — review the result before spending another credit.

Dream-versus-truth versus the full three-slide format

Both formats have three stages and use one credit in Version 1. Confession templates use two generated images and a blurred text break. Progressive templates use three generated images. The selected preview determines which structure applies. For motion rather than assembled images, choose Version 2 and check its duration and quality rates on pricing.

Sound, captions and the AI label

The video is silent on purpose, so the finish is up to you.

Add the sound. The trend audio is usually “Coming Up Roses”, credited to Harry Styles. Add it inside TikTok or Reels yourself rather than expecting the download to carry it. Adding audio through a platform does not grant you rights to use it anywhere else.

Write the caption. Name the costume the child was supposed to be. One short line is enough for the joke to land and gives the reveal its context.

Label it. Mark the dream slide as AI-generated so viewers know the costume was not real. Personal use is subject to our terms and the platform's rules.

What you receive, and what can still vary

Version 1 returns a silent, 9-second, three-stage MP4 and generated images. Your upload serves as an identity reference; the final stage is generated rather than an unchanged copy. Review each stage and the assembled video in My videos. Version 2 is a separate AI motion workflow with 10- or 15-second options and its own credit rates.

Mistakes that flatten the reveal

A few choices make the difference between a punchline and a confusing clip.

A group photo. Use a single clear subject. Group shots make the intended identity less clear.

A hidden costume. If the homemade outfit is not visible, the truth slide has nothing to reveal. Pick a photo that shows what went wrong.

A filtered face. Heavy filters remove the detail the AI needs to hold identity. Prefer an evenly lit, lightly processed image.

Rushing to post. Review all three slides before sharing, and only post when the people involved agree. Read the photo checklist for more.

The superhero: pillow muscles into a cape

The superhero look keeps one clear child photo. Pillows stuffed under a shirt become a real hero's cape. Upload one portrait; the AI redraws the costume and keeps the face and pose. Choose this when the homemade outfit is part of the joke.

The mermaid: a towel tail into a real mermaid

The mermaid look uses one childhood photo. A towel wrapped as a tail becomes a real mermaid. This is a different costume brief, not a promise of fewer artifacts; review the face throughout the result.