Deepfake Studio

Deep Works

Contains adsIn-app purchases

About this app

Create data sets to use with AI models that train and generate images and videos. Our current version allows optimization for faces. You can replace your selfies with wacky effects depending on your data set as long as there is at least one face present.

With Deepfake Studio, you can use your device to train and run inference or use our servers with powerful hardware that does the work for you.

Ratings and reviews

I absolutely love this app... Although I prefer the old version over 2.0 within the app itself. Yes, 2.0 offers you the ability to edit offline and it is much faster too. But the swap capabilities are way less than adequate. I'd much rather pay and wait for a high quality swap than get faster results that are poor at best.... However I'm not saying this to diminish the apps capabilities. It really is probably the best deepfake app on the market. Definitely better than Reface. 😉👍🏻

The quality of 'fakes' is impressive, however, this is a pay to play only, in the most insidious way. You can earn credits by watching ads, but are only allowed two lifetime ads. Making an image is free. The quality is super great, it just takes 30 minutes. Making a 5 second video is free, but the queue limit will always be full and above what allows you to add a job to do. This is where they have you. The rest costs credits, and is easy when you have them. Pay to win scheme on crack.

This app is not worth your space. You have to use credits for every single function, and you dont even start with enough to upload a single image. You need 20 credits to make one face set, which is ridiculous. I would not recommend this app, the entire credit system is an obnoxious way to get revenue with ads, or straight up purchases. Don't install

What's new

2.1.11
- replaced previous models with new xS and xM versions which are 3x more efficient, requiring significantly less training and produces much better results
- Users can now train and swap online without passing a check to train on device first
- Import existing data sets to other models
- Bug fixes