Comment on page

Technical Details

In-depth insights into ELNA's technical foundation, canister framework, and key components.

Technical Foundation

ELNA is built upon the foundation of the Internet Computer, a groundbreaking decentralized computing platform. Leveraging the unique attributes of the Internet Computer, ELNA addresses the challenges of creating decentralized AI solutions. Key technical aspects of ELNA include:

Canister Framework

At the core of ELNA's technical architecture is the Canister Framework. This framework provides the essential infrastructure for various AI-related tasks, including data ingestion, model training, bot deployment, and serving queries on-chain. It offers the following capabilities:
  • Decentralized Execution: Canisters allow for decentralized execution of AI processes, ensuring that AI functionality is distributed across the Internet Computer's network.
  • Scalability: ELNA leverages the Internet Computer's scalability to support larger and more complex AI models as the platform grows.
  • Interoperability: Canisters are interoperable with other chains and systems, allowing ELNA to integrate seamlessly with a broader ecosystem of decentralized technologies.

Vector Database

One of the key components of ELNA's technical infrastructure is the Vector Database. This specialized database is optimized for on-chain data storage and retrieval, making it a critical element for managing AI data efficiently. It plays a pivotal role in:
  • Data Storage: The Vector Database efficiently stores the data necessary for AI model training and inference on the Internet Computer.
  • Data Retrieval: ELNA's architecture allows for rapid retrieval of data from the Vector Database, enabling real-time AI interactions.

Access Control and Identity Management

ELNA prioritizes user data privacy and security through robust access control and identity management. It integrates with the Internet Identity system, ensuring that users have secure and authenticated access to ELNA's AI capabilities.


ELNA provides workflows that simplify complex AI processes. These workflows enable:
  • Custom Data Ingestion: Users and developers can easily ingest custom data sets, allowing them to train AI models on domain-specific information.
  • Model Training: ELNA orchestrates the training of AI models across canisters and, when necessary, with external inference APIs.
  • Bot Deployment: Trained models are deployed into isolated canisters, ensuring the security and integrity of AI assistants.

Conversational AI

ELNA's architecture includes specialized components for conversational AI. This capability allows ELNA to engage in dynamic and context-aware conversations, providing natural and responsive interactions for users.
In the following sections, we will delve deeper into ELNA's development phases, governance model, and strategies for community engagement and adoption.