Ungate InfiniRoute AVS
  • 🔥Introduction
    • Overview
    • Scope
  • 📓ARCHITECTURE
    • Abstract
    • Keys Components
  • 🧠Inference Engine
    • Abstract
    • User Interface Layer
    • Coordination Layer
    • Data Layer
  • ⚓AVS
    • Abstract
    • Task Consensus
    • Consensus Engine
    • Rewards and Penalties
  • 🏪Models Marketplace
    • Abstract
    • Externally Hosted Models
    • Benchmarking Engine
    • Marketplace Economy
  • 👩‍💼Operators
    • MainNet
      • Registration
      • Installation
      • Quickstart
    • TestNet
      • Installation
      • Registration
      • Quickstart
  • 🎁Summary
    • Immediate Next Steps
    • Building the First SubNet
    • Future Vision
    • Conclusion
Powered by GitBook
On this page
  1. ARCHITECTURE

Keys Components

PreviousAbstractNextAbstract

Last updated 6 months ago

The InfiniRoute system contains three main component groups

Inference Engine

The Inference Engine is the heart of the InfiniRoute system, responsible for dynamically routing AI inference requests to the most suitable models. This component is designed to optimise performance based on cost and accuracy parameters, ensuring each prompt is processed by the best available model.

AVS

The Actively Verification Service (AVS) using is a critical component for secure and transparent metadata management within the InfiniRoute architecture. It ensures that all metadata generated during inference is securely stored onchain, providing verifiable and immutable records.

Models Marketplace

The Models Marketplace has externally hosted models, specialised AI models hosted on various platforms and utilised by the Inference Engine for processing prompts. These models are fine-tuned with proprietary data. Model producers connect their models into the marketplace through APIs and that makes it accessible to the router.

📓