Aizel Network: An On-Chain Trustless AI Inference Network Supporting Large Models Aizel Network makes AI trustless. Aizel is an end-to-end on-chain verifiable AI inference infrastructure that empowers blockchain users or dapps (smart contracts) to use any trustless AI abilities. On Aizel, dapps and users who need to use trustless AI for all kinds of scenarios, including gaming, social, DeFi, etc., can all directly plug into Aizel and receive the verifiable AI inference produced in Aizel with a no-code level of setup.

Aizel persists in pre-trained machine learning models in a distributed and scalable architecture, produces and attests AI inference in a decentralized and trustless manner, combining the approach of TEE + MPC, and modularizes this verifiable inference ability to the Avax networks. Benefited from the highly scalable and top-speed remote attestation system the team develops, Aizel can attest to large models directly on-chain, meaning not only conventional AI models but also large models with billions of hyper-parameters can be provided with total trustless capacity for on-chain usage. In such a way, Aizel can massively elevate the ceiling for on-chain AI.

Currently, Aizel is the only project in the space that can execute trustless LLM/large models. The following document explains specific technical designs in more detail. https://docsend.com/view/jy74xa8dxxg7i96p