The whole process, start to finish
Making a Caught at the Window video takes one photo and three short steps. This guide walks through each one, including the two places people usually lose quality: the photo choice and the posting finish.
Step 1: Pick the photo (the step that decides everything)
The scene is fixed, so your photo is the only creative input. The generator needs to read a face clearly enough to keep it recognisable through a push-in shot and a mid-bite freeze.
Use a photo that is:
- Front-facing and well lit, with both eyes visible
- Unobstructed: no sunglasses, no hand over the mouth, no heavy filters
- Single subject: one adult, centred, with some headroom
- Recent enough to still look like the person
Wedding outfits, work uniforms and costumes are all fine: whatever the person wears in the photo is what they wear at the window table. The best photo guide goes deeper, with the failure cases to avoid.
The consent rule is simple: your own photo, or an adult's photo with their permission. No children, no celebrities.
Step 2: Generate
Sign in, open the generator, upload the photo and select Generate. One render costs 8 credits; pack options are on the pricing page. There is no prompt field: the scene, the camera move and the timing are the fixed template, so two people using the generator get the same staging with different stars.
Rendering usually takes a few minutes. You can leave the page; the finished video waits in your library.
Step 3: Check the render like a viewer
Watch it once at phone size and check three things: the face stays recognisable at the mid-bite freeze, the stare lands before the curtain moves, and the curtain fully closes the scene. Small AI variations in hands or background objects are normal and most viewers never notice them mid-scroll.
If the render failed outright, the 8 credits are already back in your balance automatically.
Step 4: Post it with the trend's sound
The exported video carries the template soundtrack. In TikTok, upload the clip and attach the current Caught at the Window sound over it, so the joke's timing rides on whatever is trending that week and the post joins the sound page. and let the app's text tools add your caption: the format's favourites are variations on "caught lacking" and "he saw the camera". The TikTok sharing guide covers sounds, captions and posting times in detail.
That is the whole workflow. Photo, generate, post: the generator is ready when you are.