Documentation
Satya AI Marketplace Documentation
Satya is a decentralized marketplace for AI models built on the SUI blockchain. It enables secure trading of machine learning models using advanced encryption, trusted execution environments, and blockchain technology.
Core Features
Decentralized Model Trading
Trade AI models through a structured pipeline: Upload → Pending → TEE Verification → Marketplace Listing. All transactions are recorded on the SUI blockchain using smart contracts for transparency and immutability. Creators maintain ownership while enabling controlled access through encrypted storage.
SEAL Homomorphic Encryption
Models are protected using SEAL encryption with threshold cryptography (2 out of 2 key servers). Encrypted models are stored on Walrus with access policies enforced through smart contracts. Session-based access keys are provided to buyers after successful purchase verification.
Trusted Execution Environment
TEE verification provides cryptographic proof of model integrity and performance. Models are executed in secure enclaves to generate attestations that buyers can verify, ensuring models perform as advertised before purchase.
Walrus Decentralized Storage
Encrypted model data is stored on Walrus testnet with blob IDs tracked in smart contracts. Maximum file size is 1GB with 10MB chunk sizes for efficient uploads. Storage includes 5-epoch default retention with aggregator and publisher endpoints.
Architecture Overview
The platform consists of several integrated components working together to provide a secure and efficient marketplace experience: