Play Jean Phil

ONE PHOTO · SIX MODELS · ONE SONG

Singing Dog AI Models Compared — Six Ways to Make Your Dog Sing From One Photo

Play Jean Phil is a singing dog AI that turns one photo of your dog or cat into a 5, 10 or 15-second singing video made by the video model you choose, from $0.25 (5 credits), 3 free credits when you sign up. Six models: Seedance 2.0 Mini is the cheapest, Seedance 2.5 the most detailed. For pet owners and creators comparing models before they pay. Videos run 480p to 2K, 9:16, 1:1 or 16:9, watermarked until a first purchase.

1–2 photos · 6 models · 5, 10 or 15 s · Your MP3/WAV or typed words · From 5 credits · MP4

Pet photo

JPG, PNG or WebP · face clearly visible

Second pet (optional)

Add one to make a duet

A sharp, front-facing photo with the mouth and nose free works best. With two photos, photo 1 sings on the left.

Scene
What it sings

The model writes the voice and the music in the same pass. Leave the words empty for a short original song.

Video model
Length
Resolution
Frame

This is exactly what the model reads. Change the scene, the outfit or the mood; @Image1 is your first photo, @Audio1 your song.

1200 / 3000

Sound is part of the video: the singing voice and the music, or your own song.

Seedance 2.0 Mini · 10 s · 480p · 9:16 · 1 pet(s)

This video costs 10 credits

About $0.50 at the Studio pack rate

Videos from accounts without a purchase carry a small moving site mark; any purchase unlocks the clean file.

Credit packs · Your photos and song are used only for this video and are not shown on the site.

Bathtub opera: a ginger cat in a towel · 10 s, 9:16

The October 2026 version of the trend: wrapped in a white towel, a hair dryer in one paw and a brush in the other, the cat closes its eyes and belts out the high notes while the dryer blows its fur.

Which AI model is best for an AI singing dog video?

There is no single best model; each one wins on one thing. Seedance 2.0 Mini is the cheapest (10 credits for 10 s at 480p) and the default. Seedance 2.5 keeps the most detail in fur and face, but costs 49 to 268 credits for 10 s and checks your photos first. Wan 3.0 is the cheapest 1080p that takes your own song. MiniMax H3 makes sharp 2K vertical clips but only sings words you type.

The table only lists what can be checked: what each model accepts, its resolutions and its price here. We have not run a paid side-by-side test on one photo, so treat a first Mini clip as your own test.

ModelYour own songResolutions10 s priceChecks the photo firstGood for
Seedance 2.0 MiniYes (MP3/WAV 2–15 s)480p, 720p10–20NoA cheap first try (default)
Seedance 2.0 FastYes480p, 720p20–43NoThe 2.0 family’s fast tier at Mini’s sizes
Seedance 2.0 ProYes480p, 720p, 1080p34–178No1080p from the 2.0 family
Seedance 2.5Yes480p, 720p, 1080p49–268Yes, a few minutesThe most detail (fur, eyes, whiskers)
MiniMax H3No, typed words only768P, 2K29–46NoSharp 2K vertical with typed words
Wan 3.0Yes480p, 720p, 1080p17–66NoThe cheapest 1080p with your own song

Which model to pick for what you want to make

Start from what you want, not from the model name. To see whether your photo works at all, a 5 s Mini clip at 480p costs 5 credits. For a clip to post with words the AI sings, Mini at 720p is often enough. For your own song in 1080p, use Wan 3.0. For the most faithful likeness, Seedance 2.5 at 720p. Then change one setting at a time.

What you wantModel and settingCredits
Test a photoSeedance 2.0 Mini · 5 s · 480p5
A clip to post, words sung by the AISeedance 2.0 Mini · 10 s · 720p20
Your own song, in 1080pWan 3.0 · 10 s · 1080p66
The most faithful likenessSeedance 2.5 · 10 s · 720p109
2K vertical, typed wordsMiniMax H3 · 10 s · 2K46

Which models can sing my own song?

Five of the six. Seedance 2.0 Mini, Fast and Pro, Seedance 2.5 and Wan 3.0 take an MP3 or WAV of 2 to 15 seconds as a reference audio, and the pet’s mouth follows the vocals in that file. MiniMax H3 takes no audio: it sings words you type (up to 160 characters) or Happy Birthday with a name. When you pick “Upload my song”, the model list only shows the five that can use it.

  • Cut the chorus or the 10 to 15 seconds you want: clear vocals, no long instrumental intro.
  • The video is set at least as long as the song (a 12 s song moves the length to 15 s).
  • A very fast rap may drift by a syllable; a slower part lines up better.
  • The length is measured on the server from the file, not only in your browser.

How much does each model cost?

A video costs from 5 credits (Seedance 2.0 Mini, 5 s, 480p, about $0.25) to 402 credits (Seedance 2.5, 15 s, 1080p). The price depends on the model, the length and the resolution, never on the frame or the number of pets. The default, Mini 10 s at 480p, is 10 credits (about $0.50). The exact price is on the button before you start.

ModelResolution5 s10 s15 s
Seedance 2.0 Mini480p51014
Seedance 2.0 Mini720p102030
Seedance 2.0 Fast480p102030
Seedance 2.0 Fast720p224365
Seedance 2.0 Pro480p173450
Seedance 2.0 Pro720p3672107
Seedance 2.0 Pro1080p89178266
Seedance 2.5480p254973
Seedance 2.5720p55109163
Seedance 2.51080p134268402
MiniMax H3768P152943
MiniMax H32K234669
Wan 3.0480p91725
Wan 3.0720p173350
Wan 3.01080p336699

One credit is about $0.050 in the Studio pack ($69.99 for 1,400 credits). Packs are one-time purchases, not subscriptions, and the credits do not expire.

QuestionHere
Free credits3 credits on sign-up, enough for 0 Mini 5 s try(s)
Cheapest video5 credits (Mini, 5 s, 480p)
Smallest pack$14.99 for 250 credits
WatermarkA small moving site mark until the account makes a purchase; a purchase also unlocks the clean copy of videos already made
Failed videoCredits come back automatically, also when Seedance 2.5 refuses a photo
DownloadMP4 with the singing in the file

Credits from the $5 Trial pack are for the Jean Phil tool on the home page only.

Can Meta AI, Kling or Runway make my dog sing?

They can animate a dog photo, but singing needs two more things: sound in the video and a mouth that follows that sound. A general image-to-video model moves the mouth when you ask, without necessarily matching a given song. Before paying, check whether it takes an audio file, whether it lip-syncs to it, and whether it keeps the animal’s muzzle instead of drawing a human mouth.

The models offered here are general video models too (Seedance, MiniMax, Wan). The difference is the prompt and the flow: the photo goes in as the identity reference, the song as the audio reference, and the prompt spells out what must stay the animal’s own. You can read and edit that prompt in the form.

ToolTakes your songLip-syncs to the audioPrice
General image-to-video mode (Meta AI, Kling, Runway)Check for each modeCheck; often built for human facesTheir own plans
CapCut “Singing Pets” templateNo, the template’s songTo the template’s songIn the CapCut app
This pageYes with five models (2–15 s)Yes: the prompt asks for the animal’s own muzzle, no human lipsFrom 5 credits

How to make your dog sing with the model you picked

Six steps in the browser, on a phone or a computer, with no app and no editing. The order starts with the cheapest model, so a photo that does not work costs little. Most videos are ready in two to five minutes; Seedance 2.5 adds a few minutes to check the photos.

  1. Add a sharp, front-facing photo

    A photo where the face looks at the camera and the mouth and nose are free. A second photo makes two pets sing a duet (photo 1 on the left).

  2. Choose a scene

    Bathtub opera, theatre stage, recording studio or Happy Birthday. The scene only changes the prompt.

  3. Type the words or add your song

    Words for the AI to sing (or the name for Happy Birthday), or an MP3/WAV of 2 to 15 s.

  4. Pick the model

    Start with Seedance 2.0 Mini (5 credits for 5 s); switch only if the first try asks for it.

  5. Set length, resolution and frame

    5, 10 or 15 s; 480p to 1080p (H3: 768P or 2K); vertical 9:16, square 1:1 or wide 16:9. Read the prompt and edit it if you like.

  6. Start it and download the MP4

    Tick the box about the photo and song rights, press the button that shows the price, then download the video.

Singing dog AI video examples

Five singing pets, one for each scene the tool makes: the bathtub opera, the birthday, the theatre stage and two studio takes. Each clip has its singing in the file; tap to play with sound. They show the kind of result, not a comparison between models.

Wrong result? What to change, and when to switch models

Change the photo or the song first, the model second. Most bad results come from a soft, side-on or covered-mouth photo, or a song that is too fast. When the input is good and the face still drifts, a higher resolution or Seedance 2.5 usually keeps more detail. A video that fails on the provider’s side is refunded.

ProblemFirstThen try
The snout or breed changesA sharp, front-on photo with the head large in frameSeedance 2.0 Pro or 2.5 at 720p
The mouth lags behind the songA slower 8–12 s part with clear vocalsA Seedance model; H3 cannot take a song
Extra paws or a human handKeep the prompt’s “no human hands” lineRun it again, or Seedance 2.5
Smeared fur, soft pictureA sharper, well-lit photo720p or 1080p instead of 480p
Very dark fur, odd teethA brighter photo with eyes and nose visibleSeedance 2.5 or Wan 3.0 at 1080p
Seedance 2.5 refuses the photoCredits come back automaticallyAnother photo, or another model

Compared with Singing Pet, Freebeat and TuneArt

The singing pet tools ChatGPT cites most do not say which video model they run; they sell one result. Here you choose from six named models with a price per model, length and resolution. Their figures below are what their own pages state; where a page gives no figure, the cell says so.

ToolVideo modelFree partOutputYour own song
Singing PetNot stated1 video a month, watermarkedUp to 30 sYes (audio upload)
FreebeatNot stated (“animal mode”)Sign-up credits1080p, native 9:16Yes (MP3/WAV)
TuneArtNot statedFirst video free, small watermarkMusic-video style clip with lyricsNot stated
This pageSix named, your choice3 credits on sign-up5–15 s, up to 1080p or 2KYes, 2–15 s (5 models)

Other tools’ figures as stated on their own pages on 2026-10-24; they may change, so check before you pay.

What it can’t do

It makes one or two pets sing from photos, and nothing more. It does not speak in a talking voice, does not copy your dog’s real bark, and cannot promise every word or name is sung exactly: the voice is made by the model.

  • No more than two pets, and no starting video (photos only).
  • Songs of 15 seconds at most; MiniMax H3 takes no song.
  • No 4K; 1080p at most (2K with H3 in vertical).
  • No app: it all runs in the browser.
  • The watermark stays until the account makes a purchase.

What is the singing dog trend?

Making pet photos sing is not new: in 2022 the Revive app already animated animal photos to songs. The big meme came in November 2024, with a toy poodle rapping in Chinese to SKAI ISYOURGOD’s “Yin Guo” (因果), the “mountains of gold and silver” part. In October 2026 the new version shows pets standing in a bathtub, wrapped in a towel, singing opera into a hair dryer.

Meme origin: Know Your Meme: Chinese Dog Singing

Singing dog AI questions

Which model is cheapest for a first test?

Seedance 2.0 Mini: 5 credits for 5 s at 480p, about $0.25. It is the default model.

Which model keeps my dog’s face most recognisable?

Seedance 2.5 keeps the most detail, especially at 720p or 1080p. The photo still matters more than the model: front-on, sharp, mouth free.

Does a higher resolution make the lip sync better?

No. It makes the picture sharper (fur, eyes), not the mouth more accurate. For the sync, pick a slower part of the song with clear vocals.

Why does Seedance 2.5 check my photo before it starts?

This model only takes real photos after the provider has checked them. It takes a few minutes; if a photo is refused, the credits come back automatically.

Can MiniMax H3 sing my uploaded song?

No. H3 only sings words you type or Happy Birthday with a name. For your own song, pick a Seedance model or Wan 3.0.

Can two pets sing in one video?

Yes: add a second photo. They take turns and then sing together, photo 1 on the left, each keeping its own fur. The price stays the same.

Is dog singing AI the same as singing pet AI or an AI singing animal tool?

Yes, they are names for the same thing: a photo of an animal is animated so its own mouth sings a song. This page works with dogs, cats and other pets with a visible face.

Is there a singing dog AI app, or do I need one?

You do not need one: this page runs in the browser on iPhone, Android and desktop, and the video downloads as an MP4.

On this site

Play Jean Phil · All AI tools · Credit packs

Model and tool names belong to their owners. Only upload photos of your own pets, or with the owner’s permission, and songs you may use.