Restore one image in a focused Studio app
Drag and drop an image here
Or tap anywhere in this card to upload. PNG, JPG and WEBP are supported.
PNG, JPG, WEBP
Output format
Safety tolerance
Seed
What the AI Image Restoration app is for
Use this workflow when the image already exists and the main job is repair, cleanup, face recovery, colorization, denoising, or deblurring rather than broader prompt-based image generation.
Why this restoration workflow is separate
Five restoration workflows in one focused app
This app groups general photo repair, face restoration, colorization, denoising, deblurring, and stereo super-resolution into one single-image workflow with dedicated model tabs.
- FLUX Restore for broad old-photo repair and colorization
- CodeFormer and GFPGAN for face restoration
- DDColor for grayscale colorization
- NAFNet for denoise, deblur, and stereo tasks
Before-and-after review in one place
Once the output is ready, the app shifts into a side-by-side comparison so you can inspect the original image and the restored result without leaving the workflow.
- Centered upload first
- Original image on one side
- Restored result on the other
Task-specific controls instead of a generic prompt form
Each model tab exposes the actual controls that matter for restoration, like fidelity, denoise task type, colorization model size, or output format.
- Schema-aligned image input
- Fixed low-cost and premium options
- Optional stereo-pair support for NAFNet
AI Image Restoration use cases
This app works best when cleanup and recovery are the main task, not a side effect of image generation.
Repair older photographs with scratches, blur, faded detail, or missing color before preservation or sharing.
Recover faces from compressed, low-quality, or AI-generated images before reuse in profiles, thumbnails, or marketing assets.
Denoise, deblur, or colorize images before moving into compositing, upscaling, background removal, or broader image editing workflows.
How to use the AI Image Restoration app
Upload one image, choose the right restoration model, then compare the original and restored output in one place.
Upload one source image
Start by uploading a single photo or image that needs repair, face restoration, colorization, denoising, or deblurring.
Choose the restoration model
Pick FLUX Restore, CodeFormer, GFPGAN, DDColor, or NAFNet above the workload, then adjust the small set of model-specific controls before you run.
Review and download the result
Compare the uploaded image with the restored output, then download the result directly from the app.
How to choose the right restoration model
Choose FLUX Restore when the image needs a broad repair pass and you want one model that can clean damage, fix scratches, and colorize older images. Choose CodeFormer or GFPGAN when facial recovery is the highest priority and the job is mostly about portraits, headshots, avatars, or people in older photos.
Choose DDColor when the image is largely grayscale and the main task is realistic colorization. Choose NAFNet when the issue is more technical, such as noise, motion blur, or stereo super-resolution, rather than face repair or historical-photo restoration.
When to stay here versus move into other image tools
Stay in AI Image Restoration when the source image is already compositionally correct and the task is to recover quality from an existing file. This app is the shortest path from damaged input to a cleaner restored output.
Move into AI Upscaler when the next task is enlargement. Move into AI Remove Background when the next task is subject isolation. Move into the broader AI Image Generator when the repaired image still needs prompt-driven editing, compositing, or larger creative changes.
Commercial-use note for CodeFormer
Replicate currently marks the CodeFormer API as non-commercial. The tab is still available so you can compare outputs inside the app, but commercial usage should follow the model's stated license and restrictions.
AI Image Restoration FAQs
Helpful answers about inputs, model fit, and credit cost for this app.
Related workflows
Open related workflows when restoration is one step inside a larger cleanup, product-photo, or image-editing process.
Create and edit images in one shared workflow across top Studio image models.
Enhance one image with Recraft, Google, and Clarity AI upscaler options.
Remove backgrounds from one image and export a clean transparent PNG in one pass.
Create cleaner ecommerce product photos with background removal, upscaling, and polished image generation.
Restore one image now
Open the workload for a focused one-image restoration flow, or move into the broader AI Image workspace if you need follow-up edits after repair.