Transcription
What if I told you some influencers aren't even real? They're completely AI generated and the people behind them are making serious money.
In this video, I'll show you exactly how to create your own AI influencer with Comfy UI and how to monetize it on Fan View. Creating an AI influencer isn't just about pretty pictures. It's about consistency, creativity, and smart workflows. In this video, you'll learn how to design a fully custom AI character, train it to stay consistent, turn still images into animated clips, and give your influencer a polished, lifelike finish. Then, at the end, we'll cover how to launch and monetize your AI persona on Fan View.
Now, before we dive in, let's quickly talk about the setup. If you're already using Comfy UI, either locally or on RunPod, you're good to go. But if you're brand new and don't have it installed yet, I've linked setup tutorials in the description that'll get you running in just a few minutes. I'll be using RunPod since it makes everything faster and smoother, especially when we get to training models and generating videos. If you're signing up for the first time, use the link in the description to get up to a $500 bonus on your first deposit.
All right, now let's load into Comfy UI and start designing our AI influencer. First, download the Flux Dev model from hugging face and drop it into your model/checkpoints folder. Then, inside ComfyUI, go to the workflow menu. Click on browse templates and load up the Fluxdev workflow. Once that's ready, it's time to generate your images. Think of how you want your influencer to look and write this in the prompt box. Give them distinctive features like unique facial traits or hairstyle. For the best results, I recommend these settings. In the empty SD3 latent image node, set width and height to 1024. And in the K sampler node, set steps at 30 and set the euler to simple or beta. With those locked in, hit run, and Comfy UI will bring your AI character to life.
Once you're happy with the image you've created, it's time to take things further by training a custom LoRA. But before we can do that, we need a solid data set, and that's where Flux Context comes in. Flux Context is great for creating variations of your character while keeping their defining features consistent. To get started, first download the Flux Context model and place it into your model/checkpoints folder so Comfy UI can recognize it. Once that's set, load the Flux context workflow into Comfy UI. Then, drag in the image you just generated as your base. From here, you'll write prompts that keep the core look of your character, like facial structure, hair, and eye color the same while changing other aspects such as expression, pose, or angle. For example, you could write, "Keep the hairstyle and facial features the same, but give her a surprised expression with wide open eyes slightly turned to the side, hands near her face." Using this method, you can quickly build a diverse set of images, close-ups, half-body shots, and full body poses, all while preserving your character's unique style. Aim for around 30 to 40 high-quality images. That's the sweet spot for training a reliable LoRA. The more consistent your character is across images, the stronger and more professional they'll look when you start posting on Fan View.
Once your data set is ready, it's time to train your own custom LoRA. Now, I won't go super in-depth here because we already have a full detailed tutorial on LoRA training. You'll find that linked below if you want the step-by-step breakdown, but let's quickly go over the essentials. We'll be using Flux Gym on RunPod to handle the LoRA training. To get set up, log in and click on pods in the sidebar. Search for an RTX 4090 and select it. Then, press change template and search for Next Diffusion and choose the Flux Gym template. Once your GPU is running, connect to port 8888 to open a VS Code server. In the terminal, type ./run_fluxgym.sh and wait for it to finish loading. After that, connect to port 7860 from runpod and flux gym will open in your browser. Just make sure to keep the VS Code server open while you're using it.
Inside Flux Gym, the first thing you'll do is give your LoRA a name and set a trigger word. This trigger word is what you'll use later in prompts to tell Comfy UI to apply your LoRA. After that, drop in all the images from your data set. Next, let Chat GPT generate captions for each image. Just make sure to include your trigger word in each caption so the model learns it properly. Copy those captions and paste them back into Flux Gym next to the corresponding image. Now, for the settings, keep your images resized to 1024. Set repeats per image to five and train for about 12 epochs. The learning rate should be set to 5e-4 and set network dim to 16. I also recommend saving every couple of epochs and setting sample images to generate every 400 steps. Just write a few short prompts so you can track how your training is progressing. That's it. Once training starts, you'll see sample images pop up showing progress. And after about 2 hours on a 4090, you'll have your very own trained LoRA ready to test.
Once your training finishes, download the LoRA files from VS Code workspace and place them inside the model/lora folder. That way, Comfy UI can recognize them. Now, let's test it. Open up the Fluxdev LoRA workflow in Comfy UI. Find the LoRA loader node and select your trained file. Don't forget to enter the trigger word you chose earlier at the start of your prompt. Once your LoRA is loaded, write a prompt similar to your training data, like a portrait or a full body scene, and generate your first images. This is the exciting part where you really see your consistent AI influencer come to life. Here's a tip. Sometimes the best results don't come from the very last epoch. Try loading earlier versions like epoch 8 or 10 and compare them to the final one. You might find one that nails the style even better.
The next step is optional, but highly recommended if you want your images to really shine. It's a blazing fast face detailer workflow, an amazing face enhancer that works in just a couple of seconds. I'll link the full tutorial in the description, but in short, it adds natural texture to the skin, sharpens the eyes, and makes the whole face look more realistic. Again, not required, but it will boost the quality of your images a lot.
To really maximize your revenue on Fan View, you don't just want to post pictures. You'll want some spicy videos, too. And the great part is you don't need to film anything. With Comfy, you can animate the images you've already made using the Wan 2.2 imagetovideo workflow. Here's how it works. First, load the WAN 2.2 workflow inside Comfy UI. Then, take one of your best images, the ones we just polished, and drop it straight into the workflow as your base. From there, you'll guide how the video should move by writing a short prompt. For example, you can write, "Slow head turn toward the camera. a natural blinking warm cinematic lighting." Once you hit run, Wan takes your still image and brings it to life by generating smooth, natural-looking motion. Depending on the settings, you'll get a short clip that feels like your influencer is really moving. The whole process is fast and straightforward, and the results are perfect for stories, posts, or premium content on Fan View. If you want the full breakdown of settings and options, I've linked a detailed tutorial in the description. This also includes all the model downloads.
Now that we've got our influencer looking great in both photos and videos, it's time to take the final step, actually putting them out into the world and starting to earn. And for that, we're heading over to Fan View.
All right, go to fanview.com and sign up. Once you're logged in, you'll see a button on the homepage that says become a creator. Click it. We can skip the first step since we don't have a referral code. Next, choose your country of residence. Make sure it matches the country listed on your identification. In this step, we'll choose between AI created or human created content. Since we're building an AI influencer, we'll go with AI created and continue. We need to decide what type of content we want to create on the platform. There's a wide variety of categories to choose from, but let's be honest, we need to spice things up. That's why we're going to select the NSFW category. After all, our AI influencer has what it takes to make it all happen. On the platform, we'll be identifying as female. So go ahead and select that option, then continue. Now we have the option to set our profile picture, display name, and handle. We'll skip this for now and add them later. The same goes for our banner and bio. We'll set these up later, so we'll skip them for now. In the next step, we can set our monthly subscription price, but we'll leave it at the default for now. Later, we can create subscription bundles in the dashboard, offering discounts for multiple months instead of just a single monthly price, giving our fans more options.
In this step, we need to verify our identity. Even for an AI influencer, this is just a standard KYC check to prevent fraud. For this identity check, you'll need to scan the QR code on the screen with your mobile phone and follow the steps shown. This process is handled by Andato, a secure third-party service used by Fan View. They follow all regulations, encrypt your personal data, and are a fully valid company, so there's no need to worry. Once your identity check is complete, you'll be redirected back to Fan View. If you mention how you heard about Fan View in this section, you'll earn 100% of your earnings for the first seven days. So why not mention Next Diffusion? Now click continue. And congratulations, you've completed the setup and are one step closer to earning money with your AI influencer.
Now that we're in Fan View, we can start working on our profile. Right now it's only 13% complete as shown in the top right corner. Let's change that. First, click on the complete your profile button. This will guide you through the steps we need to complete. We'll take our time with each one to really spice up this profile. Let's start by adding our profile picture. With the knowledge we've gained and our custom LoRA we have trained, it's time to jump back into Comfy UI to create a stunning profile picture. We'll use Fluxdev as our base model and load our LoRA, making sure to include the trigger word in the prompt along with a clear positive description for the picture. The setup is simple. a one by one ratio at 1024 x 1024 25 steps CFG set to one Euler as the sampler and beta for the Euler. That's it. Couldn't be easier. The result is a beautiful profile picture for our AI influencer. We've completed another step and are now at 25%. If you don't see a green check mark after uploading the image, simply refresh the page to update your progress.
Now it's time to create our profile banner. We'll head back into Comfy UI to make it. Let's go. We'll use the same workflow as before, but adjust the dimensions to the correct banner size, 1,200 pixels wide by 352 pixels high. Next, we'll provide a prompt and keep all other settings the same. And just like that, our banner is ready. It's a perfect eye-catching addition for our Fan View profile. As you can see, we're getting into the groove with completing our steps. Already at 38%. Great progress. Let's keep going together.
The next step is to update our bio. So, click on the next step. And here we will change the name and handle, adding a short introduction about our AI influencer, and finally updating the location. We'll begin by giving her the name Kira Lunz, a choice that matches her appearance and feels tropical. For the handle, we'll use the same name in all lowercase with a dot in between. For the bio, we want to make it engaging and spicy. The goal is to attract fans into subscribing, so don't be afraid to add a playful, slightly cheeky tone. For the location, we'll set this beauty as based in Brazil. It perfectly matches her appearance and bio. Of course, don't forget to hit save in the top right corner. Now, let's see where our progress bar leads us. We've hit the 50% mark with our AI influencer setup. Not quite finished, but we're making progress. By setting everything up correctly, we can maximize visibility and increase fan engagement. Let's keep going.
Next, we'll upload an intro video for our beauty, Kira. Click on add intro video and toggle the intro video option. I've already prepared a video using the Wan 2.2 imagetovideo model. This video will be featured for discoverability and has a high chance of attracting potential fans. Make sure it contains no nudity as it will be visible to everyone visiting the site. Now, let's take a look at the intro video of our beauty. When you have uploaded your video, make sure you click on save at the bottom of the page. Let's scroll back to the top and return to the edit page.
Our next step is to set up automated messages for new subscribers, followers, and other specific events. For now, we'll set up the basic welcome messages for new followers and subscribers so they feel right at home on my page. Fan View makes it super easy with the generate message with AI feature. It creates a message based on our profile info and bio, so most of the time it's ready to go. I've already done this for the new subscriber message. So, now let's do the same for new followers. Let's give it a try. I don't really like the first welcome message it came up with for my new followers. It feels a bit too long. I won't save it just yet. Let's try again. Click regenerate and see what it comes up with this time. After a few tweaks, I'm happy with the message to send to every new follower. Click save. And now you'll see that my new follower message is set to on. This part is all set and ready to go.
Now head back to our profile page and check our progress with the initial setup. Click on your profile in the top left corner. We're at 75%. Let's see which steps are still left to complete. Next up, creating our first post, but hold on. Since you're not subscribed, you won't get a sneak peek, so we'll skip this step for now. The final step, connect your social media. Link TikTok, Instagram, or any other platform by adding your username and follower count. You can even link directly from your other social accounts to your Fan View page to grow your fan base. This is a real game-changer.
I hope this video got you excited to start your journey on Fan View. Check out their website and our written guide to begin creating a stunning AI influencer and start earning your first dollars with your AI persona. Thanks for watching and I'll see you in the next one.