Upcoming v2.3.0: What to expect

[w4ffl35]

I have been working on polishing AI runner for the upcoming v2.3.0.

I'm currently revamping the canvas and have already optimized many other areas of the app which were causing bottlenecks by moving some processes into workers in separate threads. The performance boost is noticeable.

Many improvements to the UX have been made. The current version (2.2.0) feels like a prototype. My goal is to have v2.3.0 feeling like a more polished and mature version of the application, so I have redesigned much of the interface, simplified some tools and improved the menu systems.

AI Runner was never meant to be JUST an AI art application. v2.3.0 introduces the ability to use LLMs in a variety of ways. You will have the ability to choose from a variety of LLMs. Out of the box you will have the ability to chat with a bot similar to ChatGPT. You can use the LLM to generate prompts for Stable Diffusion which can help you with the creative process (similar to DALL-E).

With the right GPU you can achieve near real-time text to speech, and you can choose from a variety of voices and other text to speech settings (currently I have integrated Bark and SpeechT5). I am also currently working on Speech to Text.
The goal here is to control the application (and more) with voice activated LLMs, and speak to your bot in near real-time.

The overall architecture of the app has been completely reworked (an ongoing process). I am also working on a vision to text module, and a number of other features which I will eventually showcase in a video. In the meantime, consider supporting me on Patreon.