I'm building an open source alternative to Topaz Photo AI

Before I start, I know that many people, including myself, have an aversion to "AI apps" nowadays, because these projects are often just thin wrappers around ChatGPT's API to do the heavy lifting.

This is not one of those projects. In this case, all the AI logic in my project is implemented directly: data pre-processing, post-processing, inference, tiling, tensor operations, and everything in between are handled entirely within the project, with no external API doing the heavy lifting.

With that said…

This project aims to be an open-source alternative to Topaz Photo AI. For those unfamiliar with Topaz Labs, they develop paid, proprietary AI applications for enhancing images and videos. While there are free and open-source tools that can achieve similar results, such as ComfyUI, they are often difficult to use. Topaz products, on the other hand, place a strong emphasis on ease of use.

Both Topaz and ComfyUI are excellent tools. That said, I personally rely on Topaz applications most of the time because they cover 90% of my needs with minimal friction. Recently, however, Topaz Labs switched to a subscription-based licensing model, which I strongly dislike. That decision motivated me to build an open-source alternative that matches Topaz’s usability without the subscription lock-in.

My project is called Open Photo AI, completely open source and free. It currently support enhancements like AI upscaling, face recovery and light adjustment. It doesn't have yet all the features that Topaz has, but it's a starting point. My goal is to continue developing it into a fully featured alternative to Topaz Photo AI. If that succeeds, I plan to move on to alternatives for Topaz Video AI and other tools.

I’d appreciate any constructive feedback.