What is Signals & Sorcery?

Signals & Sorcery is a platform for creating, sharing, and consuming machine learning operations. A developer can write a python function, publish it, host it, and have it consumed from native applications such as Ableton, Figma, and Chrome Extensions.

Who is this for?

• Application power users who want to extend their applications with self-hosted SOTA AI operations. ex. META releases a new audio generation model and a user wants to use it in their DAW immediately.
• Developers creating Colab/Jupyter notebooks who want to use them in the context of a native application. ex. A developer creates a new audio stem splitting model and wants to use it in int the context of their DAW.

How does it work?

Signals & Sorcery is composed of three core components.

• A network discovery server. Essentially a system that brokers data transfer between the Crucible Plugins and the Rune AI's.
• A Python3 pip package runes-cli. The package is used to package, invoke and publish a python functions as Rune AI's.
• Native plugins called Crucible Plugins. The plugins are essentially embedded web-forms. When a Rune is connected to a curcible-plugin an interface this is generated from the Rune AI's function signature. The user can call the Runes function and pass data to-and-fro from the plugin to the Rune. In the case of Crucible-Audio the user can drag/drop audio/midi files from the DAW into plugin, which is then transferred to the Runes compute instance.