demobook

OpenArt: Generate a character reference image with GPT Image 2

Demo summary

The user selects the GPT Image 2 model in OpenArt, attaches the custom character, and uses a specific '@' prompt to generate a photorealistic reference image of himself as a samurai.

Step-by-step

  1. Go to the image section and select the GPT Image 2 model
  2. Navigate to the reference section and click on characters
  3. Select the custom character you previously built
  4. Enter your prompt in the prompt box, ensuring you use the @ symbol before the character's name
  5. Set the aspect ratio to 16x9
  6. Select the quality level (Low, Medium, or High)
  7. Choose the number of images to generate (up to four)
  8. Click Generate

Options

  • Select Low quality to test ideas at the cheapest price
  • Select Medium quality for a balance between speed and quality
  • Select High quality for the best possible reference image
  • Generate up to four images at once

Watch out for

  • You must use the @ symbol before the character's name in the prompt or the AI will guess the character and lose consistency

Tips

  • Use the GPT Image 2 model for the best photorealistic results
  • Use the 16x9 ratio to make the generation feel like a cinematic scene

Highlights

if you've used Nano Banana Pro before, this model feels way better

All demos from “How to Make AI Videos That Really Look Like You (Full Guide)

  1. 0:420:39Create a custom AI character in OpenArtThe user demonstrates how to build a consistent AI character by uploading personal photos to OpenArt's character section and naming the profile.OpenArtAI Avatar Generator
  2. 1:301:21Generate a character reference image with GPT Image 2CurrentThe user selects the GPT Image 2 model in OpenArt, attaches the custom character, and uses a specific '@' prompt to generate a photorealistic reference image of himself as a samurai.OpenArtAI Realistic Image Generator
  3. 4:340:42Generate consistent environment prompts using ClaudeThe user uploads reference images to Claude to extract visual cues and generate a descriptive prompt for a Japanese dojo that matches the existing cinematic style.ClaudeAI Image Generator
  4. 5:200:23Generate environment background in OpenArtThe user pastes the Claude-generated prompt into OpenArt to create a high-quality cinematic background image of a dojo.OpenArtAI Background Generator
  5. 6:220:59Animate AI video with SeaArt (Seedance 2)The user demonstrates setting up a video generation task in SeaArt by uploading character and location references and using '@' tags in the prompt to control the action.SeaArtAI Animation Generator
  6. 11:460:15Stitch AI clips in CapCutThe user demonstrates dragging the generated AI video scenes onto the CapCut timeline to assemble the final edited video.CapCutAI Video Editor
  7. Watch “How to Make AI Videos That Really Look Like You (Full Guide)” →

AI Realistic Image Generator

  1. 4:321:13Photorealistic image generation with Nano Banana ProThe user selects the Nano Banana Pro model and uses a prompt for a woman in a coffee shop to generate a high-quality, photorealistic image.Thomas Creates
  2. 4:350:40Test cinematic realism with SeaDance 2.0A detailed prompt for an elderly man in an Italian cafe is used to demonstrate SeaDance 2.0's ability to render skin textures and lighting.Affiliation
  3. 4:300:32Create hyperrealistic celebrity portraitsThe user demonstrates generating a high-detail cinematic portrait of Margot Robbie using Nano Banana 2 on OpenArt, bypassing standard Gemini restrictions.Dan Kieft
  4. 1:512:29Generate a base AI character with OpenArtThe user demonstrates how to select the 'OpenArt Photo Realistic' model and input a detailed prompt to generate the initial high-quality image of an AI influencer.Roboverse
  5. 8:060:18Photorealistic generation with OpenArt Photo RealisticThe user demonstrates OpenArt's proprietary model designed to create images that look like raw mobile phone snapshots.Roboverse
  6. 5:480:24Create a consistent location assetThe creator demonstrates generating a high-detail jungle temple interior image in OpenArt to serve as a visual anchor for the environment in subsequent video clips.Isa does AI
  7. 7:250:25Generate high-resolution 4K macro imagesThe video shows the process of selecting 4K resolution and generating a macro shot of a honey drop with high visual fidelity and air bubble details.Isa does AI
  8. 1:301:21Generate a character reference image with GPT Image 2CurrentThe user selects the GPT Image 2 model in OpenArt, attaches the custom character, and uses a specific '@' prompt to generate a photorealistic reference image of himself as a samurai.Youri van Hofwegen