How to Make AI Selfie Videos with VEO3
Easy step-by-step tutorial how to generate consistent character viral AI selfie talking videos using VEO 3 Fast
Meet me in Salt Lake City, October 11-14! Use promo code SABRINA500. I don’t earn affiliate commission or payment, just want to meet more women in AI :)
Here’s a viral AI selfie video for Tiktok Shop, 100% generated with VEO 3, which scored 6.5M views and raked in crazy sales for Father’s Day:
Tiktok failed to load.Enable 3rd party cookies or use another browser
In this tutorial, I’ll share the exact prompt template to generate consistent character AI selfie talking videos, just like above, using Google Gemini’s VEO3 Fast.
VEO3 Fast is the faster cheaper version of VEO3, so it’s particularly awesome for rapid prototyping and storyboarding.
Let’s start with my AI selfie prompt template in JSON format.
Remember, you can copy/paste this prompt into ChatGPT and have it edit the prompt with your desired specifics:
the environmental setting
the dialogue you want the AI to say
any unique visual or voice character traits
{
"subject": {
"age": 22,
"gender": "male",
"facial_features": {
"face_shape": "oval with a slightly angular chin",
"eyes": {
"color": "hazel (green with brown flecks)",
"shape": "almond-shaped, slightly hooded",
"expression": "alert and curious"
},
"hair": {
"color": "dark brown",
"length": "short",
"style": "messy, wind-tousled with a few longer front strands",
"texture": "slightly wavy"
},
"nose": "straight with a subtle bump on the bridge (likely healed break)",
"lips": "medium fullness, often pressed in a faint half-smile",
"jawline": "defined and lean, with light stubble most days",
"cheekbones": "moderately high",
"distinctive_marks": [
"thin pale scar running diagonally across the right eyebrow",
"a scatter of light freckles across the bridge of the nose",
"small faded burn mark on the left forearm (visible near wrist)"
]
},
"voice": {
"tone": "energetic and earnest",
"timbre": "clear with a slight huskiness when tired"
},
"accent": "Northern English (soft, short vowels typical of Yorkshire/Manchester region)",
"clothing_style": "weathered dark-brown traveler's cloak with a patched hood",
"accessories": [
"old leather backpack with reinforced straps and a brass buckle",
"worn leather gloves tucked into a belt",
"simple braided leather cord necklace with a small carved wooden charm",
"phone in hand (selfie-style)"
],
"overall_impression": "youthful, resourceful, and perpetually on the move"
},
"context": {
"scene": "edge of a misty cliff, looking out at the vast ocean below",
"atmosphere": "soft coastal mist, distant rolling waves, wind moving cloak and hair",
"time_of_day": "overcast daylight with diffused, soft highlights through fog"
},
"camera": {
"framing": "selfie perspective, chest-up (selfie chest-up close framing)",
"motion": "natural handheld motion with slight shake, occasional small re-framing and subtle push-in to face typical of TikTok",
"style": "realistic, TikTok realism — casual, handheld, authentic"
},
"ambiance": {
"lighting": "soft, diffused light filtered through mist; slightly cool color temperature with warm skin tones",
"tone": "intimate, cinematic yet raw and shareable for social media",
"ambient_sfx": [
"distant ocean waves",
"soft wind gusts",
"faint cliffside seagull calls"
]
},
"action": {
"behavior": "holds phone at arm's length in selfie style, looks directly into camera, lets wind tousle hair and cloak, small natural gestures (tilt head, brief smile), occasionally pans slightly to show ocean behind before returning focus to face",
"expression": "alert, curious, earnest energy"
},
"audio": {
"voice": {
"tone": "energetic and earnest, clear with slight huskiness when breathy",
"accent": "Northern English (Yorkshire/Manchester)"
},
"dialogue": "Welcome to my journey. I'll take you through some of the most breathtaking places I've ever seen."
},
"visual_style": "realistic, social-media optimized",
"elements": [
"weathered dark-brown traveler's cloak with patched hood (hood pushed back)",
"old leather backpack with reinforced straps and brass buckle visible over one shoulder",
"worn leather gloves tucked into belt",
"braided leather cord necklace with small carved wooden charm",
"thin scar across right eyebrow and light freckles on nose visible",
"small faded burn mark on left forearm near wrist (visible when framing includes forearm)",
"misty cliff edge foreground rocks and sparse windswept grass",
"vast ocean and low cloud layer in background",
"handheld phone visible in frame (selfie grip)"
]
}Using Blotato “AI Selfie” Template
Next, let’s create AI selfie videos using this prompt template.
I’ll show you 2 options: using Blotato vs. using Google Gemini.
The biggest advantages using a 3rd party app like Blotato:
mix & match VEO3 with 10+ other AI image/video models
instantly publish or schedule posts to social platforms
video editor makes it easy to assemble & edit multiple scenes
Navigate to Blotato video maker: https://my.blotato.com/videos/new
For your 1st run, you don’t need to edit the form. You can just use default values and click “Generate Video” at the bottom.
Notice that you can specify the scene for each shot, what you want the AI character to say, and you can input a character reference.
When the video is ready, you’ll see a video editor. Click on a PINK bar, which corresponds to a VEO 3 clip. You’ll see the exact prompt that was used to create it, on the lefthand side.
Here’s my AI selfie talking video, first attempt, no edits, yapping about teaching AI!
Using Gemini Google Website
Now let’s do the same thing using the Google Gemini website.
Navigate to Gemini: https://gemini.google.com
Select “Create Videos with Veo”:
Copy/paste your AI selfie prompt and wait a few minutes. Here’s my VEO 3 selfie video, after uploading my headshot as the character reference image. Amazing results, first attempt, no edits.
Need More Help? 👋
1/ Free AI courses & playbooks here
2/ Free AI prompts & AI automations
3/ I built Blotato to grow 1M+ followers in 1 year