Image
Insightface
State-of-the-art 2D and 3D Face Analysis Project
★ 29.9k starsCustom (check repo)
View guide →[NeurIPS 2022] Towards Robust Blind Face Restoration with Codebook Lookup Transformer
CodeFormer is an open-source image project by sczhou (sczhou/CodeFormer on GitHub). It has 18,171 GitHub stars and 3,725 forks, is released under the Custom (check repo) license, and was last updated on GitHub on 18 November 2025.
Generation studio, upscaler, background remover, or photo library — match the tool to the task.
Desktop apps, Docker images, and portable builds cover most platforms; pick the simplest for your OS.
Generation tools need checkpoint files; utilities like upscalers usually bundle everything.
Generate, enhance, or import — then export, share, or keep building your library.
The defining trait of this whole category: professional results without monthly fees.
Your images stay on your machine — important for client work and privacy.
Bring any compatible open checkpoint; you are never locked to one vendor's style.
Single-job tools (upscale, de-background, restore) that do one thing excellently.
Self-hosted managers with AI tagging, face recognition, and deduplication.
Extensions, scripts, and community presets extend the popular studios far beyond stock.
Download the source and move into the project folder.
git clone https://github.com/sczhou/CodeFormer.git
cd CodeFormerIsolate dependencies so nothing clashes with your other projects.
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activateInstall the requirements, then follow the README's launch command (it varies per project).
pip install -r requirements.txtFor generation: a polished studio interface. For enhancement: a dedicated upscaler. For libraries: a photo manager. The guides above narrow it down.
Generation benefits hugely from one; upscalers and photo managers are far lighter.
It is an open-source project at sczhou/CodeFormer with 3,725 forks. Check the repo's contributors and recent commits to see how active maintenance is.
Mostly yes under MIT/Apache/GPL — but verify each license, especially for client deliverables.
With the right open checkpoints, absolutely — the interface matters less than the model you load.
Community hubs like CivitAI and Hugging Face host thousands of open checkpoints.
State-of-the-art 2D and 3D Face Analysis Project
Stable Diffusion web UI
Focus on prompting and generating
Real-ESRGAN — practical blind super-resolution for real-world photos and video: restores and upscales degraded images with models trained on pure synthetic data, BSD-3-Clause
MMagic — OpenMMLab's generative-AI toolbox: diffusion, text-to-image, restoration and a model zoo.
Visual diagnostics for LoRA training — spot overfitting with face biometrics
The most-starred other Image projects in the LibreAI catalogue — the realistic shortlist if CodeFormer is not the right fit.
Stable Diffusion web UI
The node-graph powerhouse for precise, professional AI image and video generation.
High performance self-hosted photo and video management solution
Focus on prompting and generating
🆙 Upscayl - #1 Free and Open Source AI Image Upscaler for Linux, MacOS and Windows
Free and open — grab the source, star the repo, and start creating.