Purpose
Krea 3 is an image-to-image AI image editor for designers, online sellers, marketers, and photographers. Its positioning is not to generate images from a blank canvas, but to edit photos users are already satisfied with: changing the lighting, color, background, or a single object while keeping faces and products recognizable. The official website emphasizes that editing runs on established models such as GPT Image 2, Nano Banana Pro, and Seedream, allowing users to use them while waiting for Krea's upcoming editing model. It should be specially noted that the official website explicitly states that this site is not affiliated with Krea AI and does not provide Krea's own model weights; if Krea opens its new editing model to third parties in the future, the site plans to evaluate integration, but does not promise a specific timeline.
Main Capabilities
Image-to-Image Editing
Upload a photo and enter a modification description, and the model will redraw according to the instructions while keeping the rest of the composition basically unchanged. Typical editing directions given by the official website include:
- Relighting: change the time of day or the direction of the light source, such as low-angle evening sunlight, a softbox on the left, or street neon, while keeping the person, pose, and composition unchanged.
- Keep the same face: upload a portrait as a reference and describe a new scene; features, hairline, and skin tone are guided by the reference image, while clothing, pose, and background can change.
- Merge multiple reference images: you can attach a handbag, a model photo, and a location photo at the same time, and explain each image's role in the prompt, for example, "put the bag from image 1 on the model from image 2." The official website states that Nano Banana 2, Seedream, and Wan 2.7 Image all support using multiple reference images in one request.
Multi-Model Image Generation
You can switch among GPT Image 2, Nano Banana Pro, Nano Banana 2, Nano Banana, Seedream 5.0 Lite, Seedream 4.5, Grok Imagine, Wan 2.7 Image, and Qwen Image 2 in the same menu. You can generate from plain text or start from your own photo, and compare how different models interpret the same prompt.
Image Editing Details
After uploading a photo, describe the modification: rearrange products, replace colors, clean up a cluttered background, or restore a faded photo. The official website explains that editing instructions apply only to the named parts, the subject remains anchored to the uploaded photo, and a slider comparison feature is provided.
AI Video
You can switch the prompt bar to AI Video, addAdd a still frame as the starting image and describe the motion. Optional Veo 3.1 Fast, Veo 3.1 Quality, Sora 2, or Seedance 1.5 Pro, supporting landscape 16:9 and portrait 9:16.
Style Presets
The official website says it offers more than 600 styles, such as oil painting, watercolor, cel-shaded anime, simulated film, and more, which can be layered on top of prompts and can be used for both text-to-image and photo editing.
My Creations Library
Every generated image and video is saved to the account, categorized by image editing, text-to-image, and video. You can view the model, aspect ratio, and resolution of any result, copy the prompt for reuse, or re-download the file at any time.
Project Finishing Tools
- 4K export and commercial use: GPT Image 2, Nano Banana Pro, Nano Banana 2, and Wan 2.7 Image support 1K, 2K, and 4K rendering. Paid exports are watermark-free; the official website says all annual plans, Studio and Max monthly plans, and Professional-tier and above bundles can be used for commercial projects.
- Background and scene replacement: keep the product or person and replace everything behind them, such as a marble countertop, a noon beach, or a city street at night.
- Multiple versions from a single prompt: Wan 2.7 Image returns up to four images at once, and most models let you set the aspect ratio first, with square for feeds, 9:16 for stories, and 16:9 for banners.
How to Use: A Four-Step Editing Workflow
- Upload a photo: You can drag in a product image, portrait, or room photo. JPG, PNG, and WebP are supported, and the photo becomes the starting point rather than a blank canvas.
- Describe the changes: Write editing instructions as if briefing a retoucher, for example "warmer light," "marble countertop," or "navy blue jacket," and you can also add style presets.
- Choose a model and generate: Choose GPT Image 2, Nano Banana Pro, Seedream, or another model; the official website says the credit price is shown before generation.
- Download or keep editing: Download the result, or upload it again as a reference for the next round. Every run is saved in My Creations.
Use Cases
The official website lists six types of teams:
- Online sellers and e-commerce: Turn one packaging image into lifestyle scenes, seasonal backgrounds, and clean studio versions while keeping the product itself recognizable.
- Marketing and advertising teams: Crop approved visual drafts into square, vertical, and banner formats, and test new background or lighting atmospheres.
- Photographers and retouchers: Relight according to text instructions, restore faded photos or remove distractions, then finish in the original editor.
- Content creators and social media: Use phone snapshots to create thumbnails, story covers, and post images, with the same face throughout the entire series.
- Designers and illustrators: Explore art directions with style presets, combine reference images to make mood boards, and give clients multiple options on the same morning.
- Real estate and interior staging: Redecorate vacant or outdated rooms, change daylight to dusk, and show buyers options before moving in furniture.
Billing Methods
The official website provides two billing paths:
Subscription (billed annually, total price displayed, renewed annually until canceled)
- Starter: 800 credits/month (9600 credits/year), up to 80 images/month.
- Pro: 2000 credits/month (24000 credits/year), up to 200 images/month.
- Studio: 6000 credits/month (72000 credits/year), up to 600 images/month.
- Max: 14000 credits/month (168000 credits/year), up to 1400 images/month.
One-time purchase (credits valid for 10 years from purchase)
- Starter Package: up to 100 images.
- Growth Package: up to 300 images.
- Professional Package: up to 1000 images.
- Business Package: up to 2000 images.
- Agency Package: up to 3200 images.
- Enterprise Package: up to 7000 images.
New accounts can get free credits for a trial. A single image is billed by model and resolution, and the official website states the price is 10 to 80 credits.
Limitations and Notes
- This site is not affiliated with Krea AI and does not provide Krea's own model weights; it currently uses third-party models.
- Commercial rights are provided only with paid options that explicitly list a Commercial Use License: all annual plans, the Studio and Max monthly plans, and the Professional, Business, Agency, and Enterprise packages. Users need to own or obtain permission to use uploadedphotos.
- Uploads support JPG, PNG, and WebP; output sizes depend on the model. GPT Image 2, Nano Banana Pro, Nano Banana 2, and Wan 2.7 Image provide 1K, 2K, and 4K presets, while other models return standard resolutions.
- Editing effects such as relighting or colorization vary by model, and the official website recommends keeping multiple outputs for comparison first.
- Regarding information such as the release time and general availability of the Krea 3 model itself, the official website states that it was announced on August 28, 2026, as a next-generation foundation model based on Krea 2, but it was not yet generally available at the time of the announcement; this model is currently not included in this editor.
