NemoClaw Knowledge Wiki

Gemini Omni AI: How I AI's Video Avatar Cloning Experiment Report

4 min read

Gemini Omni AI: How I AI’s Video Avatar Cloning Experiment Report

Generated: 2026-07-24 · API: Gemini 2.5 Flash · Modes: Summary

Gemini Omni AI: How I AI’s Video Avatar Cloning Experiment Report

Clip title: I cloned myself with Gemini Omni in 15 minutes (and it’s terrifyingly good) Author / channel: How I AI URL: https://www.youtube.com/watch?v=UNZczH0gpHc

Summary

The video showcases Claire Ho, host of the “How I AI” podcast, attempting to create an AI-powered video avatar of herself using Google Flow and the new Gemini Omni video generation model. Her primary goal was to quickly produce a minute-long promotional video for her podcast using her AI likeness, a task she admits to having struggled with in previous attempts. She framed the episode as an experimental adventure, uncertain if the ambitious endeavor would succeed.

The process began with Claire generating her own avatar within Google Flow by scanning a QR code with her phone and following a series of prompts, including speaking a sequence of numbers and performing head movements. Once the avatar was successfully created (albeit with a slight “fish-eye lens” effect), she prompted Google Flow to develop a storyboard for a “hype video” for her podcast. Claire provided specific stylistic and contextual details, envisioning a “dark home office” with “dark green walls,” “books about AI and fun posters,” and a “hacker vibe” that was both authentic and high-tech.

Google Flow responded with a detailed seven-frame storyboard, outlining various scenes from extreme close-ups of typing hands to a “host reveal” of Claire, a “workflow montage,” and a “lifestyle/authentic moment.” Although the AI initially generated static images, Claire corrected the settings to produce dynamic video clips, incorporating elements from her real office background that the AI intelligently recognized from her avatar selfie. While the generated videos captured the essence of the requested scenes, Claire noted instances of “uncanny valley” and some stylistic inconsistencies, such as varying backgrounds or slightly distorted facial expressions across different frames.

Despite these minor imperfections, Claire successfully stitched together the various AI-generated video clips, complete with her voice narrating the storyboard text, into a cohesive minute-long promotional video in a remarkably short timeframe of approximately 10-15 minutes. Her key takeaway was the immense power of these multimodal generative AI tools to democratize complex creative tasks, allowing individuals without specialized video production skills to quickly bring their ideas to life. The experiment highlighted the impressive potential of AI to accelerate content creation and unlock new avenues for digital expression, even if the results occasionally ventured into the humorous side of artificial intelligence.

Video Description & Links

Description

In this experimental episode, I document my real-time attempt to create an AI avatar of myself using Google Flow and the new Gemini Omni video generation model. I walk through the entire process—from scanning my face with my phone to generating a complete one-minute hype video for the podcast, all in about 15 minutes.

What you’ll learn:

  1. How to create an AI avatar using Google Flow in under five minutes
  2. Why video AI tools unlock creative possibilities for people with zero video production skills
  3. The step-by-step process of generating a full storyboard using AI as your creative producer
  4. How to use character consistency features to generate multiple video scenes with the same avatar
  5. The uncanny-valley moments you’ll encounter when your AI clone doesn’t quite nail emotions or physics
  6. How to stitch together AI-generated scenes into a complete video using built-in editing tools

Brought to you by: Merge—Connective infrastructure for production AI: https://www.merge.dev/howiai Jira Product Discovery—Prioritize with insights, build with confidence: https://atlassian.com/howiai

In this episode, we cover: (00:00) Getting started with Google Flow and Gemini Omni (01:38) The avatar creation process: scanning and photo capture (02:55) Using Flow to brainstorm a hype video storyboard (06:59) Generating the first video scene with the avatar (08:41) Troubleshooting: accidentally generating images instead of videos (09:32) Generating all seven scenes for the complete video (11:37) Reviewing the avatar videos (13:13) Stitching the videos together in the browser-based editor (14:32) The complete How I AI hype video (15:32) What worked and what didn’t (19:04) Final thoughts

Blog & detailed workflow walkthroughs from this episode: How I Built an AI Avatar and Hype Video in 15 Minutes with Google Flow: https://www.chatprd.ai/how-i-ai/ai-avatar-video-in-15-minutes-with-google-omni-flow ↳ How to Create a Promotional Video with an AI Creative Director: https://www.chatprd.ai/how-i-ai/workflows/how-to-create-a-promotional-video-with-an-ai-creative-director ↳ How to Create a Personalized AI Avatar with Google Flow: https://www.chatprd.ai/how-i-ai/workflows/how-to-create-a-personalized-ai-avatar-with-google-flow

Tools referenced: • Google Flow: https://labs.google/fx/tools/flow • Gemini Omni: https://gemini.google/overview/video-generation/ • Veo 3: https://deepmind.google/technologies/veo/

Where to find Claire Vo: ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo

Production and marketing by https://penname.co/. For inquiries about sponsoring the podcast, email jordan@penname.co.

URLs

Graph View