Hello,
If anyone is working on AI integration, please reach out via PM if you see any of the following as a good 'match' for your project.
I have been working on an AI back-end framework for modular synths and it is now ready for BETA testing.
Please know it's not a complete solution, only the critical parts of the back-end and hardware framework required to implement AI. ( the parts needed no matter what AI solution type is implemented, like packet protocols, module addressing, interface to 'train' the model, saving/loading module settings to/from cloud, javascript node API etc )
It's difficult to summarize but here is a partial list of what I have finished so far. I'm focused on clock and sequencer AI to start. ( find good clock/polyrythm/sequence combos and share them with others )
- Web browser/iOS/iPAD app. ( allow user to 'like' or train the AI as part of the control surface )
- Protocol to 'chain' modules together and auto-address via 3 wire cable. ( automatically tell the AI what devices are on deck )
- Load/save module settings to localstore/eeprom/cloud via node/javascript. ( provide framework for any integration )
- Communicate with an AI server back end, passing the data required to do AI back and forth.
- Auto firmware updates from browser. ( encrypted bootloader )
- Packet protocol to communicate between browser/cloud and hardware modules. ( this has been my main focus for several years now )
- Full API on a dedicated 'STM32'. Can be 'dropped into' existing project or 3rd party modules like an ASIC
If sees this as a good match for their project, PM me and we can discuss in detail. It's quite a large project for just 1 person :)
The api guide can be found here: https://roboDNA.com