What This Is
You’ve probably seen those viral videos where someone is filming a selfie and a wild animal casually walks up to them. A polar bear, a panther, a lion. The person reacts, the animal snuggles them, and it looks completely real.
These are made with AI video generators. You upload a reference photo of yourself, paste a detailed prompt, and the AI creates the video. The prompts below are built to keep your face consistent throughout the clip while adding a realistic animal interaction.
What Tool to Use
These prompts run on Google Veo 3.1, which you can access through Freepik or AIVideo. Both platforms have Veo 3.1 built in, so you don’t need a separate Google account or API access. Just pick one, upload your reference photo, and paste the prompt.
1/ Polar Bear Prompt
This prompt creates an 8-second selfie video set in a luxury mansion living room. You’re filming yourself, talking to camera, and a polar bear slowly walks in from behind the couch. It approaches you, rubs against your cheek, and licks your face. You react naturally.
@aitrendz.xyz You don’t even need to be on camera anymore in 2026 Just clone yourself with AI and you can create ANY super realistic scene 🙌 Here’s the prompt I used (copy paste it from me for free)👇 Polar Bear prompt: Ultra realistic 8-second selfie video of the SAME man from the reference image, filmed inside a luxury mansion living room (marble columns, ornate white ceiling, chandelier, grey walls, couch behind). Selfie mode close to face, slightly wide lens, natural indoor daylight from front, crisp face focus + shallow depth of field. He smiles and says: “Hey cherry! Come here I want people to see how beautiful you are! Good boy…” Background stays identical (NO scene change). From behind the couch, a large realistic polar bear slowly walks into frame, tail swaying, soft paw steps, sniffing air. It approaches the man’s shoulder. He notices and laughs nervously. polar bear rubs against his cheek like a cat, softly growls/snorts, then gently licks his face. He laughs and says: “Aww good boy…” Extreme realism: visible pores, stubble, micro skin texture + shadows, realistic leopard fur with individual hairs/spots, whiskers, wet nose, realistic tongue licking. Ultra sharp 8K, 64K micro-detail, clean natural grading, no flicker/warping/face change
♬ original sound – René Remsik
Here’s the full prompt:
Setup: Ultra realistic 8-second selfie video of the SAME man from the reference image, filmed inside a luxury mansion living room (marble columns, ornate white ceiling, chandelier, grey walls, couch behind). Selfie mode close to face, slightly wide lens, natural indoor daylight from front, crisp face focus + shallow depth of field.
Dialogue: He smiles and says: “Hey cherry! Come here I want people to see how beautiful you are! Good boy…”
Action: Background stays identical (NO scene change). From behind the couch, a large realistic polar bear slowly walks into frame, tail swaying, soft paw steps, sniffing air. It approaches the man’s shoulder. He notices and laughs nervously. Polar bear rubs against his cheek like a cat, softly growls/snorts, then gently licks his face. He laughs and says: “Aww good boy…”
Realism details: Extreme realism: visible pores, stubble, micro skin texture + shadows, realistic fur with individual hairs, whiskers, wet nose, realistic tongue licking. Ultra sharp 8K, 64K micro-detail, clean natural grading, no flicker/warping/face change.
2/ Black Panther Prompt
Same concept, same setting, but with a black panther instead. The animal behavior is slightly different since panthers move differently, but the structure is the same.
Here’s the full prompt:
Setup: Ultra realistic 8-second selfie video of the SAME man from the reference image, filmed inside a luxury mansion living room (marble columns, ornate white ceiling, chandelier, grey walls, couch behind). Selfie mode close to face, slightly wide lens, natural indoor daylight from front, crisp face focus + shallow depth of field.
Dialogue: He smiles and says: “Hey cherry! Come here I want people to see how beautiful you are! Good boy…”
Action: Background stays identical (NO scene change). From behind the couch, a large realistic black panther slowly walks into frame, tail swaying, soft paw steps, sniffing air. It approaches the man’s shoulder. He notices and laughs nervously. Black panther rubs against his cheek like a cat, softly growls/snorts, then gently licks his face. He laughs and says: “Aww good boy…”
Realism details: Extreme realism: visible pores, stubble, micro skin texture + shadows, realistic leopard fur with individual hairs/spots, whiskers, wet nose, realistic tongue licking. Ultra sharp 8K, 64K micro-detail, clean natural grading, no flicker/warping/face change.
3/ Epstein, Donald and Lollipop Prompt
Same concept, a bit different scene but very similar results.
Here we generated the video with this JSON prompt:
{ “duration”: 8, “aspect_ratio”: “16:9”, “style”: “ultra realistic, natural, handheld camera”, “camera”: { “type”: “shoulder level handheld”, “movement”: “slight natural micro-shake, small autofocus adjustments”, “lens”: “35mm natural perspective” }, “lighting”: { “source”: “real afternoon sunlight”, “look”: “soft shadows, natural exposure, no cinematic color grading” }, “audio”: { “ambient”: “light wind, distant birds, subtle garden sounds”, “music”: “none” }, “scene”: { “location”: “realistic countryside house backyard with lawn, flowers, stone house exterior, open white doors”, “time_of_day”: “late afternoon” }, “action_sequence”: [ { “time”: “0-2”, “description”: “Man sitting casually in a wicker garden chair. Relaxed posture. He looks slightly excited, like he is about to call pets. Natural facial movement. Not posing.” }, { “time”: “2-5”, “dialogue”: “Hey Epstein, Donald, and lollipop… come out, don’t be afraid!”, “performance”: “spoken in a real, slightly excited and playful tone, like calling animals in real life”, “background_action”: [ “As he says the first name, bushes behind him move naturally.”, “A large white Yeti slowly walks out from behind garden plants. Realistic weight and fur movement.”, “When he says the second name, a huge colorful parrot flies down and lands on the bench.”, “On ‘lollipop’, a saber-tooth tiger walks out from near a tree. Grass bends under its paws.” ] }, { “time”: “5-8”, “dialogue”: “We are one big family now.”, “performance”: “he smiles naturally, slightly amused”, “description”: “Animals stand calmly behind him. Camera slightly reframes to include them all. No dramatic reaction.” } ], “realism_constraints”: [ “No cinematic slow motion”, “No fantasy glow or dramatic lighting”, “Natural animal movement physics”, “Realistic fur and feather detail”, “No exaggerated expressions”, “No artificial color grading” ] }
Here’s a YouTube video that explains how JSON prompts work and you can also grab this JSON Video Prompting Blueprint if you want to learn how this works in detail.
How to Use These
- Pick an AI video generator that supports reference images
- Upload a clear photo of your face
- Paste the prompt for the animal you want
- Generate and review the output
Tips for Better Results
The realism section at the end of each prompt matters a lot. Lines like “no flicker/warping/face change” tell the AI to keep your face stable throughout the video. Without these, AI video tools tend to distort faces mid-clip.
If the animal doesn’t look right on the first try, regenerate. AI video is still hit or miss, and sometimes it takes a few attempts to get a clean result.
You can also swap the animal for anything else. Replace “polar bear” with “tiger,” “lion,” or “wolf” and adjust the fur description accordingly. The rest of the prompt structure stays the same.
If you want to swap the setting, change the mansion description to whatever location you prefer. Just keep the “background stays identical (NO scene change)” line so the AI doesn’t randomly switch environments mid-video.
Want to find more AI tools like these? Check our homepage. Or visit the Blog for more tutorials and prompts.
