Title

"No Audio Recording Needed!" Master 100% Photorealistic Viral Videos with Google FLOW

Introduction

Hello, creators! Welcome to a2set's AI Tutorial.

When you scroll through TikTok or Instagram Reels these days, you often see natural, UGC-style videos of influencers reviewing beauty products or sharing their daily lives from a cozy room. Have you ever thought, "I want to promote our brand with that natural vibe, but hiring a pretty model every time is too expensive..."?

Today, we reveal the ultimate viral video master workflow using Google's powerful next-generation AI creation platform, 'Google FLOW', which will solve this problem in just 5 minutes.

The awkward Uncanny Valley is completely over. We will forge a flawless virtual influencer using FLOW's top-tier photorealistic rendering model, 'Nano Banana Pro'.

The most amazing part is that 'no audio recording file is needed at all'! If you just type the entire script into the video generation model, 'Veo 3.1', the AI generates a perfect native-sounding voice and matches the lip-sync flawlessly.

Moreover, don't worry if the video cuts off in the middle due to a long script. By using the '>> Extend' button and relaying the next part of the script in the prompt, you can seamlessly stitch together long-form videos.

Shall we start this incredible magic right now?

Step 1: Casting - Generating a Hyper-Realistic K-Pop Style Influencer 📸

The first step to a perfect video is creating a 'real-human-like original image' that will serve as the skeleton of the video. Google FLOW's image generator, powered by the 'Nano Banana Pro' model, is currently unmatched in expressing skin texture, peach fuzz, and natural light blooming.

1. Access your Google FLOW workspace and navigate to the image generation menu.
   
   

2. Click the model selector on the right side of the bottom prompt input box and choose the Nano Banana Pro model, which boasts overwhelming photorealistic quality.
   
   

3. Set the aspect ratio to 9:16, perfect for Reels/Shorts.
   
   

4. Copy and paste the content below into the prompt box and hit the generate button.

[Pro Prompt for Viral Influencers (For Copy & Paste)]

A raw, candid selfie video still of a beautiful K-pop style Korean female influencer sitting in a cozy, warmly lit living room. She is holding a glowing skincare serum bottle naturally. Looking directly at the camera with a bright, attractive smile. Shot on iPhone 15 Pro, front camera, UGC style, soft natural indoor lighting, photorealistic, hyper-detailed skin texture, no makeup filter, slight motion blur in the background.

💡 A2SET Tip: Using keywords like raw, candid, Shot on iPhone 15 Pro, and UGC style is crucial. It breathes life into the image, making it look like a friendly, natural daily vlog shot on a phone in your room, rather than a fake studio photo!

Step 2: Action & Voice (Veo 3.1) - Rendering the First Cut with a Dialogue Prompt 🎙️

Once the photo is perfectly prepared, it's time to make this influencer look at the camera and speak. In the past, you had to record your own voice, but now, everything is done just by typing text.

We will meticulously write down the entire dialogue we want the influencer to say in the prompt box.

1. Check the image list you just generated in the gallery.
   
   

2. Click the [three dots (More)] button on the top right of your favorite image thumbnail and select the [Add to Prompt] menu.
   
   

3. Set the video rendering model to the powerful video engine, Veo 3.1 - Fast (or Lite). Keep the ratio at 9:16 and increase the generation length generously to x4.
   
   

4. In the prompt box asking "What do you want to create?", write the mood of the video along with the 'entire dialogue text the influencer will speak'.
   
   

   • Prompt Example: >
     
     Create a realistic selfie-style video of a beautiful K-pop style Korean woman in a cozy living room. She holds the camera naturally at arm’s length, speaking directly to the viewer with a bright, confident vibe. She is speaking the dialogue: "Hi everyone! Welcome back to my channel. Today, I'm super excited to share my absolute holy grail skincare secret with you all. I've been using this glowing serum every single morning and night, and honestly, the texture is just out of this world."

5. Once you finish writing the prompt, press the Generate button! The AI will recognize the text and start rendering a natural English voice and lip-sync cut.

Step 3: The Magic Expansion - Saving Cut-Off Dialogue to the Last Lip Movement ✂️🎥

Now, this is where many creators panic: "Uh oh? I definitely wrote the entire dialogue in the prompt box, but when I check the rendered video, it cuts off abruptly after '...secret with you all.'!" This is natural. There is a physical limit (about 5-10 seconds) to the video length the AI can render at one time.

When this happens, you just need to click FLOW's magic button, '>> Expansion'. However, if you carelessly just type the next line of dialogue here, the expansion will be very awkward, and the screen will glitch. Make sure to follow the '3 Pro Tips for a Natural Extension' below!

1. Click the [>> Expansion] button at the bottom of the first video preview window that abruptly cut off at 'with you all.'!
   
   

2. When you click Extend, the AI locks the last frame (face angle, background, lighting) of the previously rendered video as the starting point for the next video.
   
   

3. [Crucial Step] Now it's time to write the remaining dialogue in the newly appeared prompt input box. NEVER just write the dialogue alone! The AI will forget the previous camera angle and atmosphere. You must copy and keep the existing style prompt, and only replace the dialogue content after dialogue: with the text that follows the cut-off point.
   
   

   • Naturally Flowing Extend Prompt Example: >
     
     Create a realistic selfie-style video of a beautiful K-pop style Korean woman in a cozy living room. She holds the camera naturally at arm’s length, speaking directly to the viewer with a bright, confident vibe. She is speaking the dialogue: "I've been using this glowing serum every single morning and night, and honestly, the texture is just out of this world. It melts right into the skin without feeling sticky."

4. Press the Generate button again, and the AI will perfectly inherit the background and lighting from right after "with you all." and render a second video speaking very naturally from "I've been using..."!

💡 Pro Tip: How to Fix Awkward Expansions!

No matter how well you write the prompt, due to the nature of Video AI, the transition to the extended cut might feel slightly glitchy or the lip-sync might seem swallowed. In such cases, apply these 2 practical techniques:

• Tip 1: Adjust the rendering length to cut at a 'Breath (Punctuation)'
  
  

  • If the video cuts off in the middle of a word (e.g., "Today I'm super exci-"), the next expansion will inevitably be awkward.
    
    

  • Adjust the video length option (x2, x4, etc.) or trim the sentence length to encourage the 1st rendering to end exactly at a period (.) or a comma (,) where the speaker takes a breath. Splicing an extension while the mouth is closed is 100 times more natural.
    
    

• Tip 2: Trim 0.5 seconds at the joint in CapCut (Magic Cut Editing)
  
  

  • By nature, the last 0.5 seconds of an AI video rendering often slightly slows down or distorts (morphing phenomenon).
    
    

  • Load the generated video clips into CapCut, and very precisely cut off the last 0.5 seconds of the front video and the first 0.5 seconds of the back video (extended video) and join them together. A flawless, smooth long-form video will be completed, so perfect that it's hard to notice the cut!

Conclusion

Let's summarize the Google FLOW master workflow for creating the most perfect influencer video:

1. [Nano Banana Pro] Extract a 100% photorealistic image down to the peach fuzz.
   
   

2. [Add to Prompt] Pass the image to the video source, and run the rendering in the [Veo 3.1] model by writing the entire dialogue prompt without uploading audio.
   
   

3. If the video cuts off in the middle, click the [>> Extend] button. Make sure to keep the existing background/style prompt, append the remaining dialogue, and refine the 0.5-second joint in your editing program for a perfect finish.

The entire process that used to require thousands of dollars for 'Planning ➡️ Casting ➡️ Studio Shooting ➡️ English Audio Mixing' is now completed in just 5 minutes with only keyboard typing, without a single microphone.

Stop worrying about expensive model casting fees or foreign language voice-overs for global expansion. Using the Master FLOW and powerful 'Extend Relay' technique you learned today, create your own exclusive global influencer without worrying about script length, and actively use it to promote your brand!