SuiSense AI
AI integrated
The problem SuiSense AI solves
SuiSense AI addresses the lack of intelligent, automated tooling for on-chain analysis, smart contract automation, and NFT metadata enrichment in the Sui ecosystem. Current DeFi and NFT platforms are limited in their ability to assess contract risk, interpret transaction behavior, or dynamically update asset metadata. By integrating AI models directly with on-chain data, SuiSense offers a powerful intelligence layer that enables:
Risk scoring for contracts and wallets using behavioral heuristics and AI classification models.
Smart contract automation powered by natural language prompts that translate to executable Sui transactions.
Dynamic NFT metadata enrichment that evolves based on market trends, on-chain activity, or user behavior.
This empowers developers, users, and platforms with better safety, transparency, and functionality without needing deep technical or data science expertise.
Challenges I ran into
One of the main challenges was connecting AI inference workflows with the Sui blockchain’s object-based data model. Unlike account-based chains, Sui’s model required restructuring how metadata was queried and stored. To overcome this, we built a custom data pipeline using the Sui SDK and RPC endpoints to index and format on-chain data for AI consumption.
Another challenge was translating natural language into smart contract actions reliably. We iterated on prompt engineering and created templates that restrict user input to predefined action types, ensuring valid and safe transaction generation.
We also faced difficulty in scaling AI inference affordably for on-chain use cases, so we designed a hybrid model using local inference for metadata and off-chain AI APIs for contract automation, preserving cost efficiency without sacrificing capability.
Tracks Applied (2)
AI
DeFi
Technologies used
Cheer Project
Cheering for a project means supporting a project you like with as little as 0.0025 ETH. Right now, you can Cheer using ETH on Arbitrum, Optimism and Base.