Pixo

Purpose

Pixo is a Raindance AI video generation tool used to combine two separate photos into a two-person performance video set on a seaside pier. Users do not need to shoot the scene themselves or write prompts; the template already provides the scene, camera moves, and reference performance, and the generated MP4 video comes with a reference audio track.

Main capabilities

• Supports uploading two separate photos, corresponding to two characters respectively: photo 1 replaces the person sitting on the pier at the beginning, and photo 2 replaces the person who walks into frame and appears in most of the close-ups.
• Selectable models: Seedance 2.5 or MiniMax H3. Seedance 2.5 offers 480p, 720p, and 1080p, up to 30 seconds; MiniMax H3 offers 720p, up to 15 seconds, and consumes fewer credits. The default settings are Seedance 2.5, 10 seconds, 720p.
• Selectable durations: 5, 10, and 15 seconds (both models), with 30 seconds limited to Seedance 2.5. All videos are in 4:3 aspect ratio, generated in a single pass, not stitched together.
• The finished video includes the audio track of the selected reference clip, with no need to find additional audio files or align them in an editor.
• The total number of credits required is shown before generation; uploading photos or changing settings will not start a paid generation.
• Supports JPG, PNG, and WebP formats, each no larger than 20 MB, with the short side at least 300 px and an image aspect ratio of 0.4–2.5.
• Image requirements: it is recommended that each image contain only one clear person; for the close-up character, choose a photo where the eyes, hair, and facial details are visible, and a half-body or full-body shot can provide more clothing and proportion information. Avoid faces with hands blocking them, dark sunglasses, or heavy filters.
• You must use photos of yourself or of participants who have consented.

How to use

1. Upload one photo for each character, check the character labels next to the preview, and swap the photos if necessary.
2. Choose the duration (5, 10, 15 seconds, with 30 seconds available for Seedance 2.5), model, and available resolution, and view the corresponding preview and total credit count.
3. Play the finished video with sound, check the faces of both people in the close-ups and shared shots, then download the MP4.

The system provides preview templates that show the first 10 seconds or the full duration of the selected length, with the user's photos replacing the performers in them.

Applicable scenarios

• Couple's two-person video: use photos taken separately by each person to place both people in the same seaside scene.
• Friends' video: specify who sits and who walks, and send it to each other after creating it.
• New attempts with the same template: replace a pair of approvedAuthorized photos, preserving the reference performance.
• Suitable for users who want to participate in the Raindance AI trend but do not want to shoot the scene themselves.
• There are three remake examples made by Tyler in the community (29 seconds), but they are not results generated by this website and are for reference only.

Limitations

• The template is fixed at two characters, and one photo must be provided for each; there is no single-person mode, and the output is 4:3 rather than native 9:16.
• Complete restoration of appearance is not guaranteed: face, hair, hands, and clothing may change, especially during turns, wide-angle shots, and close-up transitions.
• Mouth movements may differ from the reference, so the final result must be checked with sound.
• The download includes the audio track referenced by the selected template; it will not generate a new song or clone the user's voice.
• Uploaded photos will not be added to the public examples on the page; generation history and downloads require an account.
• Videos from free accounts carry a dynamic watermark; any purchase unlocks saved original videos without a watermark, including videos already made.
• If generation fails, the deducted credits will be refunded.
• New accounts currently receive 5 credits, which may not be enough to cover videos for all models; sign-up credits expire after 7 days.
• 5-second videos start at 49 credits; the specific total credits depend on duration, model, and resolution, and will be shown before generation.

Credit Packages (as provided)

• 240 credits/month: can make about 2 10-second MiniMax H3 720p videos, watermark-free videos (including those already made), and credits are refunded for failed videos.
• 780 credits/month: can make about 8 10-second MiniMax H3 720p videos, supports various durations up to 15 seconds, with priority support.
• 2,650 credits/month: can make about 27 10-second MiniMax H3 720p videos, includes all Pro features, dedicated support, and custom integrations.

Other Information

The tool is an independent video tool and is not affiliated with the artist or the creator of the reference clip. For details, refer to the Raindance AI guide, trend explanation, and step-by-step tutorial on the site.