Home / Tools / DeOldify

DeOldify

Deep learning colorizer for old photos and videos.

DeOldify preview
github.com Open Source

Upvotes

0

Pricing

Open Source

Category

Video Enhancers

In the database since

Oct 2026

Overview

About DeOldify

DeOldify is an open-source deep learning project that adds color to black-and-white photos and videos. You feed it an old image or film clip, and the model predicts realistic colors for every part of the scene. It also cleans up scratches, dust, and flicker in old footage. It suits archivists, families, and creators restoring history. The project code is free under the MIT license. Note that the GitHub repo is archived, so development has paused, but the models still work. Check the repo for the notebook and setup guides.

Key Features

1. Photo Colorization

Upload a black-and-white photo and get a color version back. The model was trained on millions of images to guess natural colors.

2. Video Colorization

Colorize old film clips frame by frame. The model keeps colors steady across frames so the video does not flicker.

3. Scratch and Damage Repair

Old photos often have scratches and spots. DeOldify can clean many of these marks while it adds color.

4. Realistic Color Model

The model learns how the real world looks. Skin, sky, grass, and clothing get believable colors without manual painting.

5. Free MIT License

The code is free for any use, including commercial projects. Download it, run it, and change it as you like.

6. Colab Notebooks

You can try it in a free cloud notebook first. No install is needed to test a photo before running it locally.

Frequently Asked Questions (FAQ)

Can DeOldify colorize videos?

Yes. It processes video frame by frame and includes fixes for flicker and jitter, which are common in old films.

Is DeOldify still maintained?

No. The repository is archived, so active development has stopped. The released models and notebooks still work, though.

How accurate are the colors?

They are smart guesses, not the original colors. Results look natural for common subjects like skin and sky, but fine details can be off.

Does it remove scratches?

Yes, partly. The video pipeline includes artifact repair that cleans scratches, dust, and noise along with colorization.

What hardware do I need?

A GPU helps a lot. You can also run it in a free cloud notebook, which needs no local hardware at all.

Conclusion

DeOldify is the classic choice for colorizing old photos and films. It is free, open source, and produces natural-looking results with little effort. Development has stopped since the repo is archived, so do not expect updates. For one-time restoration jobs, it remains one of the easiest free options. Families and archivists get good value from its simple notebooks.

Newsletter

Tools like this change fast

New Video Enhancers tools launch every week — the newsletter keeps you ahead of what's worth trying.

Keep exploring

Similar AI tools

Ranked by the job they do, their capabilities, and how closely they fit this workflow.