Repository Structure

Our codebase is organized into distinct repositories, each serving a critical function in the Kognys ecosystem:

• kognys-frontend: This repository houses the user-facing application of Kognys. It's built with modern web technologies like React and TypeScript, leveraging Vite for a fast development experience and Tailwind CSS with shadcn-ui for a responsive and intuitive user interface. This ensures a seamless and efficient experience for researchers interacting with our platform.

• kognys-agents-python: This repository contains the core implementation of our AI agents. It defines the logic for how our agents operate, reason, and collaborate within the "Chain of Debate." This repository also exposes an API that the kognys-frontend interacts with to initiate research tasks and retrieve results.

• kognys-membaseaip-api: This repository provides the Python implementation and API for interacting with Unibase Membase and AIP. It serves as the crucial link for our AI agents to record their on-chain identities, store their persistent, decentralized memories, and log every step of their "Chain of Debate" discussions on the BNB Testnet. The kognys-agents-python repository directly interfaces with this service to manage agent memory and ensure data integrity and auditability.

• kognys-da: This repository contains the configuration to deploy a Unibase DA storage node that is used by kognys-agents-python to storage the results of the Kognys research.