Skip to content

concept

aikyo is an innovative framework designed to facilitate equally balanced conversations between humans and multiple AI companions in an environment where they coexist. While recent advances in LLMs have led to numerous AI Agent implementations and A2A protocols, most focus on task automation or efficiency improvements. aikyo differs by prioritizing not operational efficiency, but rather creating AI companions that can genuinely engage with humans in a mutually respectful, equal conversational partnership.


1. P2P-Based Fully Decentralized Agent Communication

Section titled “1. P2P-Based Fully Decentralized Agent Communication”

The foundation of aikyo lies in peer-to-peer (P2P) communication. This architecture ensures each connected AI companion operates completely independently, free from reliance on centralized servers. Instead, they establish direct connections and collaborate through direct peer-to-peer interaction.

2. Dialogue Control for Equal A2A Interactions with Humans

Section titled “2. Dialogue Control for Equal A2A Interactions with Humans”

To enable natural conversation flow among multiple AI companions, the system provides the following functionalities:

  • Turn Management Implements mechanisms to facilitate smooth dialogue exchanges between LLMs. This ensures proper turn-taking and conversational sequence, creating an intuitive flow of interaction that mirrors human conversation.
  • Conversation Duplication Detection A detection feature designed to prevent multiple AI companions from simultaneously speaking the same content or engaging in unnecessary repetition. This helps maintain conversation quality and deliver more refined A2A communication.

Beyond basic conversational capabilities, aikyo offers extension features to endow AI companions with both physical embodiment and dynamic knowledge acquisition. These enable users to easily create AI companions equipped with some form of tangible presence or interactive capability.

  • Provides functionality to give AI companions operable physical forms.
  • This allows the AI companion to perform specific actions within both physical and virtual environments.
  • Enables AI companions to acquire not just pre-trained knowledge, but also dynamically obtained information that can be incorporated into their conversations and actions.
  • Through integration with external data sources or API connections, it ensures access to always up-to-date and contextually relevant information.
  • Allows the system to send requests to client applications.
  • For instance, it can transmit camera footage from the client side to the AI companion or enable the AI to use voice synthesis capabilities available on the client side at any desired time.