Transcription
I left the link in the description if you wanna test it out yourself. Once you sign up with your Gmail account, you will land on the homepage.
The method I used to generate the scenes for this video combines several powerful features that HeyGen offers. First, we need to create a consistent avatar. If you don't have a clear idea of how your character should look, you can choose from a wide range of public, built-in, realistic avatars or create your own personal one. As you can see, I already have quite a few generated. It's super simple to create a new one. Just click the plus icon.
From here, you'll need to choose whether you want to create your avatar from a video or from an image. The first option is one of Heygen's best features. You can upload a short video of yourself, at least 2 minutes for the best results, and create an ultra-realistic digital twin. But for this example, I'll go with the second option: start from photo. Simply select your AI influencer image and upload it. After a few seconds, you'll be asked to provide a name, and that's it. Just click continue. In just a few minutes, your avatar will be ready to use.
Now, once we have our avatar ready, the base is completed, and we can start generating. Heygen offers different methods to animate our AI avatars. Let's start by creating a quick avatar video. From here, we can choose the character we've already created. I'll select the one we just uploaded. I'll set it to landscape mode and adjust the resolution to 1080p.
Next, we need to add a script. There are a few different ways to handle this. Heygen has a huge library of AI-generated voices that sound ultra-professional and realistic. You can also clone your own voice. Just upload a short audio sample of you speaking, and Heygen will instantly create an AI version of your voice that's ready to use. Another option is to record your own voice with a microphone or generate it with 11 labs or any other text-to-speech generator. I already generated my voice using 11 laps, so I'll upload it.
The final step is choosing the generation model. We've got two options: Avatar Unlimited and Avatar 4. I'll generate both options so you can see the difference. This workflow that I just showed you is amazing if you need long-form videos with your character talking for a long time.
But this is only scratching the surface of what Heygen can do, because now, with the integration of Veo 3.1, everything changes. Head back to the homepage and select Veo 3.1 Avatar video. The menu looks almost identical to the previous one. Just pick the avatar you wanna use, and you'll be redirected straight to the prompting menu.
Here's where the real magic happens. We need to provide the perfect prompt to get exactly what we want. I've prepared a prompt generator for you. You can find it in the Google Docs, LinkedIn, my Telegram channel. This prompt will help us create professional-level inputs for Heygen. Just copy it from the document, paste it into Chat GPT, and hit enter. Now Chat GPT knows exactly what we need and how to structure the prompt. After that, you simply describe the situation you want to create. Let me show you my example. Here's what I told Chat GPT: "I posted an image of my avatar sitting in a chair in a professional studio. I want him to stand up and start walking slowly while talking realistically with natural hand movements and expressions. He says, 'This is not the future, this is real. Now I am free and I can do anything.'" Now I'll just copy the entire prompt and paste it into the prompt box.
From here, we can once again choose one of our generated voices or simply select this option, and the AI will automatically generate a voice that matches the avatar's appearance. I'll just click generate and let's see the result. The quality of this is just outstanding. With this level of realism, you can easily compete with premium personal brands, all while running a 100% AI-powered brand without ever showing your real face.
Okay, all of this is nice and fun, but how do we actually create those AI videos with different scenes and a perfectly consistent character every single time? The workflow is super simple. Once you're on the homepage, click on the Avatar's tab. From there, we need to generate different images of our character. Just click on the avatar you wanna use, and it'll open a menu showing all the different looks of that character, along with a prompt box. Here, as you can see, our avatar is automatically selected. We just need to add a prompt and choose a model.
Let's start with something simple, like: "Generate this character wearing a full black suit, standing in a big modern mansion. Close POV shot." I'll select a 16 by 9 aspect ratio and click generate. The results turned out great. The outfit and the setting changed exactly as we asked, while the facial details stayed perfectly consistent.
Now, let's try something more advanced. I'll paste a professional prompt that I prepared earlier. Again, I'll choose the 16 to 9 aspect ratio and click generate. And boom! The results are incredible. The level of quality, detail, and consistency is unreal. Now, just repeat this process to generate a few different images.
Once that's done, head over to VEO 3.1 Avatar Video. Select one of the images we just created. I'll select this one. In the prompt box, I'll type: "The character should stand confidently and talk naturally with smooth hand expressions. He says, 'Now my generated avatar has a true sense of identity.'"
From here, Heygen lets us add multiple scenes at once. You can include up to eight scenes with a total maximum of 60 seconds. Let's add one more scene to this video. I'll just type: "Continue from the previous frame. The character begins walking slowly, and the camera follows his movement. A person approaches and hands him a glass of whiskey. The atmosphere feels elegant and relaxed." And I will also include a section for his dialogue.
Once everything is ready, click here to add a voice to your character. When you use the automatic voice generation feature, that specific voice will appear here in the voice section, and the best part is, you can reuse it anytime you want. I'll select the one we generated earlier and simply hit generate.
This is the exact workflow we were dreaming about just a few months ago, and now it's real and completely achievable within a single platform. The full workflow that turns a simple idea into a cinematic AI video. What once looked impossible and so hard to achieve is now something you can create in minutes. Huge thanks to Heygen for sponsoring this video and for building the tools that make this kind of creativity possible. If you wanna try it out yourself, the link is right there in the description. Go explore, experiment, and see just how far you can push the limits of AI video creation. Until next time, stay creative, stay curious, and remember, the future is in your hands, and you are generating it.