How many photos do I need?

Exactly two: one listener, one speaker. One-photo mode is not available.

What resolution should I pick?

720p for posting. 480p if you are testing the casting; 1080p only if you need a larger master.

How long does it take?

Usually 5–15 minutes; the timeout is 30 minutes, after which the job fails and refunds.

Can I retry a failed render?

Yes, from the result page. A retry is a new render and costs the same credits; the failed attempt was refunded.

Direct answer

Upload two clear photos — photo 1 is the listener on the left, photo 2 is the speaker on the right — pick 480p, 720p or 1080p, and render. The site returns a 15-second, 16:9, 360-frame MP4 with the original scene soundtrack mixed in. A failed render refunds the credits.

Make my video

How to make the video

The full workflow, with the exact photo order, resolutions, credit costs and download steps.

Published November 10, 2026

Before you start

You need two people, one photo each, and permission from both. The scene has exactly two seats. One photo cannot fill both; using the same image twice produces a duplicated face and a failed render.

Step 1 — choose the right photos

1Listener photo: a clear head-and-shoulders shot, face forward, eyes open, no sunglasses. This person will wear the headphones on the left.
2Speaker photo: the same quality, facing the camera or a slight three-quarter angle. This person holds the folders on the right and says the line.
3Avoid group photos, heavy beauty filters, masks, and photos where the face is smaller than a thumbnail.
4JPG, PNG or WebP; 100 KB to 10 MB per photo.

The order is the casting. If you swap the two uploads, the people swap roles. There is no “fix it after” switch; the prompt names each person by position.

Step 2 — pick a resolution

480p — 7 credits. A cheap draft for checking the casting and the timing.
720p — 16 credits. The default and the one to post; 1280 × 720.
1080p — 26 credits. Upscaled from the 720p master; useful for YouTube or client review.

Step 3 — render

The job creates one video task. The take renders, gets trimmed to 360 frames and mixed with the 15-second soundtrack. Most renders finish in 5–15 minutes. If the task stalls, the job retries; after 30 minutes it fails and refunds.

Step 4 — download

The result page plays the video and offers a download button. On desktop you can also scan the QR to open the file on your phone. Every render stays in My creations for 14 days.

Step 5 — post it

Download the MP4 and upload it to TikTok, Reels, Shorts or YouTube. For a vertical post, crop to 9:16 in the app; keep both faces centred. The file already carries the scene soundtrack, so mute it if you add the platform trend sound. Label the output as AI-generated.