Architecture Overview
To understand Unbody’s architecture, it’s important to first understand its philosophy—and why it exists.
We created Unbody to address a fundamental problem: the fragmented nature of existing tools and frameworks. These tools were designed to add AI as a feature to existing products, not to build truly AI-native products. Our goal was to build a tool that could enable developers—especially those from product backgrounds—to create AI-native products effortlessly and without needing specialized AI expertise.
We define AI-native as products that don’t just use AI but are designed to understand humans, respond intelligently, and adapt to user needs. Achieving this required us to rethink software architecture entirely, centering it around two principles:
- Knowledge Flow: Transforming raw data into actionable knowledge, not just managing static information.
- Modularity: Making every step of the process adaptable and extensible, so developers can shape workflows to meet their needs.
With these principles in mind, Unbody’s architecture is organized into four layers: Backend, API, Frontend, and Utility. Together, they form a unified stack for building AI-native applications.
This intro ties in the philosophy and rationale behind Unbody while naturally leading into the description of the architectural layers. Let me know how this feels!