I’m developing some pretty extensive AI software and I wanted to share a small little branch off my main AI project I called ManuBot.

Some of you may have seen the previous version I had named KORG Muse, but given that this AI can be trained on everything from factory maintenance of machines to human resource management of staff, ManuBot seems more fitting for what flavour of the AI I’m going to post here on the forum.
ManuBot is an offline AI Chatbot / technical support assistant that can help you get the most out of any hardware the AI is trained on. You simply ask it a question and it will help you achieve what you are trying to accomplish. Give it’s entirely offline, you can use it as much as you want without getting kicked off like other AI services, and what you said to the AI is completely private.

So, what does this AI know in general?
Well, it knows vast amount of general knowledge. You ask it to write song lyrics, or if you neeed cooking instructions on how to boil an egg, it’s got you covered. From a fined tuned musical instrument point of view and the purpose of posting it here, it has had deep training on these specific products.
ASM Hydrasynth Deluxe.
KORG Nautilus
KORG PA5X
KORG Pa4X
KORG Pa1000
Polyend Tracker Mini
Training it on more products mainly comes down to finding good data to train the AI on. Training the AI on the manuals for the product above was useful, but not entirely ideal. It is better to various sources, so the AI will be better talking about some products than others since the AI was trained on video’s and other tutorials where possible for certain products.
System Requirements.
PC:
i3 or faster.
GPU with 6.4GB of free VRAM.
Mac:
Soon...could do a built right now, but I’ve no Mac to verify my work on that platform right now.
AI Performance Information.
Running Artificial Intelligence offline locally on a computer requires a dedicated GPU with 6.5GB of VRAM free for the AI. If you have this, the AI will type back messages at a rate of 30 to 70 words per second depending on your GPU.
If you don’t have enough VRAM, it may use as much VRAM as you have, and offload the rest to system RAM. This will result in much slower response times. If you have a potato PC, then the AI may not even load.
Download Here (v0.3 Beta)
I’d appreciate any feedback you have. It early days, so I expect there to be issues. Remember, it’s a conversation you are having with an AI. So be clear on what you want it to do for you and chat with it. If you don’t speak English as your first language, tell the AI to only respond in your language and see how that goes for you.
Kind regards
Sharp