AI Couple Dance Video Generator: Turn Two Photos into One Dance
The VdoBloom AI Couple Dance Video Generator builds a partner dance video from one photo or two. Upload a picture of each person, or a single frame that already has both of them in it, choose a dance, and the AI animates the pair moving together with joined hands, a turn and a hold. Faces, hair and clothing carry over from your uploads, so the clip reads as the two of you dancing rather than two strangers who happen to look similar.
Two-photo mode is the part most people come for. Partners living in different cities, a grandparent nobody ever filmed dancing, friends who have somehow never been photographed together — the AI composes both people into one scene and matches light direction, scale and perspective so the pairing holds up. Three choreography templates ship with the tool: Waltz for an elegant ballroom hold, Salsa for something faster and warmer, and Slow Dance for a close, quiet first-dance feel. Each loads an editable prompt, so the template is a starting sentence rather than a fixed track.
The photo count also decides which models you can pick. Only the dedicated effect endpoint accepts a second image, so in two-photo mode the picker narrows to 17 model variants across five families: Runway, Kling, PixVerse, Vidu and LTX-2 Fast. Drop back to a single photo and it widens to 44 variants across 15 families, which is where the cheaper ByteDance engines live. If one of those wider-roster models is selected when you switch to two photos, the tab resets your choice to Runway Gen-3 rather than quietly dropping the second person's photo — that reset is the tool protecting your credits, not a glitch.
The arithmetic is worth stating exactly. A new account arrives with 10 free credits, and the cheapest two-photo render is Runway Gen-3 at 10 credits for five seconds at 720p — also this tab's default — so the signup grant covers precisely one couple dance clip with nothing left over. What free buys is a real generation you can watch in your dashboard, played back with a watermark over it while your account is on the free tier; a paid plan removes the watermark and unlocks the download. Past that, the prompt is yours to rewrite, duration, quality and aspect ratio are all settings, and rendering happens on our servers so you can close the tab.
How it works
- 1
Choose one photo or two
Set the photo count to match what you actually have. Two-photo mode takes a separate image of each person; single-photo mode takes one shot that already contains the couple and usually holds likeness a little tighter.
- 2
Upload the photos
Both faces should be visible and reasonably lit. Half-body or full-body shots beat tight headshots, because a partner dance needs arms and torsos to work from and the model cannot invent a hold it has never seen.
- 3
Pick the dance
Waltz, Salsa or Slow Dance. Each one sets the choreography, how close the hold is, and the lighting mood before you touch a word of prompt.
- 4
Write the scene, if you want to
Edit the prompt to name the venue, the era, the outfits or the camera move — a ballroom, a rooftop at dusk, a kitchen at two in the morning. The templates are a starting sentence, not a ceiling.
- 5
Set model and format
Choose from the models the current photo count allows, then set duration, quality and aspect ratio. Use 9:16 for Reels and TikTok, or 16:9 if the clip will play on a screen at a reception.
- 6
Generate and collect
Start the render and let it run in the background. The finished clip lands in your dashboard, where you can play it back and download it once your account is on a paid plan.
Why use this tool
Two people from two photos
Upload each partner separately and the AI composes them into a single scene, matching light direction and scale so the couple looks like they were filmed together in one room rather than pasted side by side.
Three choreography presets
Waltz for a formal ballroom hold, Salsa for a faster and more playful tempo, Slow Dance for a close, unhurried embrace. One upload can give you all three if you have the credits for them.
Faces you actually recognise
Generation is constrained to preserve facial features, hairstyle, clothing and proportions from your uploads, which is the difference between a keepsake and a generic AI couple who vaguely resemble the people you love.
A prompt you can rewrite
The text behind each template is editable, so the setting, lighting, outfits and camera language are all under your direction rather than fixed by the preset you clicked.
A picker that follows the photo count
Two-photo mode offers 17 model variants across Runway, Kling, PixVerse, Vidu and LTX-2 Fast; single-photo mode opens up to 44 variants across 15 families, and an incompatible selection is reset rather than failing mid-render.
Renders that finish without you
Generation runs server-side and the result is stored on your account, so you can start a clip, close the browser and come back to it later without regenerating.
Frequently asked questions
Is the AI couple dance generator free?
A new VdoBloom account starts with 10 credits and nothing to pay. In two-photo mode the cheapest render is Runway Gen-3 at 10 credits for five seconds at 720p, so the free grant covers exactly one couple dance clip and leaves nothing behind — we would rather give you that number than imply there is change left over. While your account is on the free tier the player shows a watermark and the download button is locked; a paid plan removes both.
Do the finished videos have a watermark?
On the free tier, yes. Free-tier playback carries a watermark overlay, and downloading, right-click saving and picture-in-picture are disabled — the clip lives in your dashboard where you can watch it. Any paid plan removes the watermark and unlocks the MP4. That distinction matters here, because couple dance clips almost always get cut into something larger such as a wedding montage or an anniversary post, and you need the actual file to do that.
Do I need to sign up to use it?
Yes, a free account is required, and it is fairer to say so plainly than to bury it at the download step. Signing up takes an email address or a Google login and no payment details. The reason is practical: the 10 free credits and your finished videos live on an account, and generation runs server-side, so there has to be somewhere for the clip to arrive.
Can I make a couple dance video from two separate photos?
Yes, and it is the main reason to use this tool rather than a general image-to-video model. The tab opens in two-photo mode: upload one image of each person and the AI places both into a single generated scene with matched lighting and perspective. If you already have a picture of the two of them together, switch the count to one photo — that path usually holds likeness a little tighter and gives you a wider choice of models.
What kind of photos work best?
Sharp, front-facing images where both faces are unobstructed. Sunglasses, heavy shadow across half a face, motion blur and very low resolution are the usual causes of a disappointing render. Prefer half-body or full-body framing over tight headshots, because a partner dance is built out of arms, shoulders and torsos — if the model can only see two heads it has to invent everything below the collar. When you are combining two separate photos, pictures shot in broadly similar light blend more easily than an indoor flash shot paired with a bright outdoor portrait.
How long does a couple dance video take to generate?
Usually a few minutes. The exact time depends on the model you selected, the duration, the resolution and how busy the queue is at that moment. You do not have to watch it happen — the render continues on our servers, so you can close the page and the finished clip will be waiting in your dashboard when you come back.
Which model should I choose for two people?
Start on Runway Gen-3, which is the tab default and the cheapest option available in two-photo mode at 10 credits for five seconds at 720p. If hands or limbs come out awkward, PixVerse v3.5 and Vidu Q2 Turbo are both 18 credits for the same clip and often read a joined hold differently. Runway Gen-4 Turbo is 20, Kling V2.5 Turbo is 40 and Runway Gen-4.5 is 48 — spend up only once you know the composition works. The cheaper ByteDance engines, including V1 Pro Fast at 7 credits and Seedance 1.5 Pro at 8, are single-photo mode only; selecting one and then switching to two photos resets you to Runway Gen-3, because only the native effect path can accept the second person's image.
Is it safe to upload photos of me and my partner?
Uploads are used to generate your video, and the results sit in your own account rather than a public gallery. Only upload photos of adults who have agreed to appear in the clip, and that includes your partner — making a dance video of someone who has not consented is not what the tool is for. Photos of minors must never be used with this tool. Generations are moderated and anything that breaches VdoBloom's content policies can be blocked or removed.
What do I do if the generation fails or the couple looks wrong?
Run it again first — these models are non-deterministic and a second attempt on identical inputs often lands cleanly. If it keeps failing, change one variable at a time: a clearer photo, a shorter clip, a lower resolution, or a different model family. When two separate photos refuse to blend, the fastest fix is usually to crop both to a similar framing so the AI is not reconciling a headshot with a full-length shot. If credits left your balance for a video that never appeared, contact support from the dashboard.
Can I post the clip on Instagram, TikTok or in a wedding video?
Yes, once you are on a paid plan and can download the file. Generate at 9:16 for Reels, TikTok and Shorts, or 16:9 if the clip is going into a slideshow that plays on a television at the reception. The download is a standard MP4, so it imports straight into CapCut, Premiere or your phone's editor if you want to add music, captions or a crossfade before it goes out.