kitio.aiAdded Nov 5, 2026
CodeFormer
PaidOpen-source face restoration model for old and AI-generated photos
What is CodeFormer?
CodeFormer is a face restoration model from Shangchen Zhou and co-authors that repairs blurry, damaged or low-quality faces in old photographs and AI-generated images. This Replicate page lets you run it in the browser or call it through an API without setting up a GPU.
The model comes from the NeurIPS 2022 paper on blind face restoration with a codebook lookup transformer. On Replicate you upload an image in the playground or send it through the API, and a prediction usually finishes within a few seconds on an Nvidia L40S GPU.
The code is public on GitHub, so you can also run it locally with Docker or try the Colab demo. It is released under the S-Lab License 1.0, which limits use to non-commercial purposes, and the Replicate listing states that its API version cannot be used commercially.
Features
- Face restoration for old photos
- Fixes faces in AI-generated images
- Browser playground
- Replicate API
- Local Docker setup
Tags
Releases
+ Submit new release- Listed
CodeFormer was added to the directory.
Organization
Shangchen Zhou (sczhou)
Listed by kitio.ai · unverified owner
Pricing
- Pricing model
- Paid
- Paid options from
- Not listed
CodeFormer plans
Pay per run
About $0.0056 / run
- Billed by Replicate per prediction
- Runs on Nvidia L40S GPUs
- Cost varies with input
Prices in USD, checked Nov 5, 2026. Plans change often, so confirm on the CodeFormer site.Official pricing
Sponsored
Your AI tool here
A Featured launch puts your tool in this spot and on the home page for a week.
Top CodeFormer alternatives
View allImage
Open-source API that turns a URL into an AI-generated image
Image
Open-source model for restoring blurry and old faces in photos
Image
Google's Gemini-based models for generating and editing images
Compare CodeFormer
If you liked CodeFormer
Run open models locally with one command, plus optional cloud models
Freemium · $20/moModels & APIs2 views3 clicks to the websiteMade in United StatesRun and deploy open-source AI models through a simple API
PaidModels & APIs1 viewsMade in United StatesFast inference and fine-tuning platform for open models
PaidModels & APIsMade in United StatesAPI that turns websites into clean markdown for AI apps
Freemium · $19/moData & analyticsMade in United StatesLow-cost serverless API for hundreds of open-source models
PaidModels & APIs1 views1 clicks to the websiteMade in United States
Pros and cons
Pros
- Runs in a few seconds per image
- Costs about half a cent per run
- Code can be run locally for free
Cons
- Non-commercial license only
- Focused on faces, not whole-image repair
- Hosted version bills per run
Prompts & results
No prompts yet. Add your own prompt and output to help others use this tool.
Share a prompt
Reviews
Q&A
What is CodeFormer?
CodeFormer is a face restoration model from Shangchen Zhou and co-authors that repairs blurry, damaged or low-quality faces in old photographs and AI-generated images. This Replicate page lets you run it in the browser or call it through an API without setting up a GPU.
Can I use CodeFormer results commercially?
No. CodeFormer uses the S-Lab License 1.0 for non-commercial use, and the Replicate listing says its API cannot be used commercially.
Can I run CodeFormer on my own computer?
Yes. The model is open source on GitHub and can be run locally with Docker, or tried in a Colab notebook.
How much does CodeFormer cost?
CodeFormer is paid.
Is there a free version of CodeFormer?
No. CodeFormer is a paid tool.
What are the main features of CodeFormer?
Face restoration for old photos, Fixes faces in AI-generated images, Browser playground, Replicate API, Local Docker setup.
Who makes CodeFormer?
CodeFormer is made by Shangchen Zhou (sczhou).
What are the best alternatives to CodeFormer?
Popular alternatives include Pollinations.AI, GFPGAN, withoutBG.