It's My Life Dog AI：一张狗狗照片制作副驾唱歌梗

@Image2 is an original three-panel visual guide, read LEFT to RIGHT: open-mouth performance, closed-mouth pause, sunglasses and resumed performance. Use it for the car setting and these ordered prop/mouth states only. Ignore the illustrated dog's breed, face and coat; replace that dog with @Image1 throughout. Use the split-screen proportions specified below rather than the guide's crop.
These panels are states of ONE subject over time, not three subjects. Synthesize continuous motion; never show the guide image, panel borders, stills or a slideshow. The sample video displayed on the page is NOT a model input.
Create a new 15-second photorealistic vertical 9:16 comedy video with generated audio from @Image1, the ONLY dog identity source. Preserve that dog's breed, coat pattern, muzzle, ears and body size. Ignore the photo background. Never turn the dog into a person. Keep the same blue collar and properly fitted pet safety harness throughout.
Use a stable horizontal split screen throughout: the upper 40 percent shows a forward-facing windshield view of a sunny road; the lower 60 percent shows the same dog sitting in the passenger seat, chest and face readable. The dog never drives. The car interior, seat, harness and daylight stay constant. The road moves gently forward without a cut; no steering interaction is shown.
0–4.5s: the dog makes three small open-close jaw movements as if enthusiastically howling, head lifting slightly. Natural canine mouth, not a human mouth or visible words. No sunglasses anywhere yet.
4.5–5.5s: the dog turns its head about 20 degrees toward the center of the car and CLOSES its mouth.
5.5–10s: hold the closed-mouth deadpan sideways look for a clearly readable pause, with only breathing and one blink. Do not keep howling during the pause.
10–11.5s: a single adult handler's hand enters only the lower panel from screen-left, gently places ONE pair of black sunglasses over the dog's eyes, then withdraws completely. The dog stays seated and does not grow human hands. Glasses must travel from the hand onto the face, never materialize.
11.5–15s: sunglasses remain correctly worn; the dog turns a little toward camera and resumes two small open-close howling movements. Finish mid-performance, same dog, same seat.
Real moving footage, not a pan or zoom across the uploaded still and not a slideshow. No panel swaps, extra dogs, duplicate glasses, human teeth, captions, logos or watermarks. Include the original audio directed below. Mouth performance is approximate and is not synchronization to any song.
Generate an original synchronized soundtrack. Low continuous car-road hum throughout, with a quiet playful ORIGINAL plucked-bass and handclap groove, never a known song. 0–4.5s: three short natural dog howls aligned approximately with the visible jaw movements. 4.5–10s: STOP the howls and musical groove for the deadpan pause; only low road hum remains. 10–11.5s: a tiny cloth rustle as the hand fits the glasses, then a soft original percussive pickup. 11.5–15s: the playful groove and two short canine howls return. No spoken words, human singing, lyrics, quotation of a melody, narration or famous recording. Keep the ear shape and hanging position from @Image1 through the glasses action. Audio and visual timing are approximate, not exact song lip synchronization.
Ear anatomy lock: Keep the uploaded dog's ear length and natural resting position through the sunglasses action. Do not borrow the guide dog's pointed ears, spread the ears sideways, or change their shape. Fit the glasses gently without lifting the ears.

狗狗在汽车副驾张嘴表演，突然闭嘴冷脸停顿，戴上墨镜后继续。常见形式是公路在上、狗狗在下的分屏。本页从照片重新生成这种画面结构。

可以，一张清晰照片提供身份，提示词要求保留品种、毛色、口鼻和耳朵，但 AI 仍可能改动。本配方只覆盖狗狗，其他动物尚未验证。

准备照片和编辑提示词免费。本 15 秒配方生成消耗 20 Raindance 积分。可用访客结账，请保管好私人恢复码，以便恢复访客积分和视频；Google 账号使用独立钱包。完成的视频正常扣费，确认失败的任务退回积分。

不含原曲。此预设生成原创器乐、狗狗嚎叫和路噪，不含 Bon Jovi 的录音或歌词。声画时序与手部交互可能变化，不保证精确口型同步。

上方视频是使用虚构主体制作的原创 AI 生成示例。分镜图是独立的可选模型输入。一个示例不能保证你的照片具有相同的动作、节奏或相似度。