FileThetic
A marketplace for AI training data you can actually trace back to its source.
ETHGlobal 1st
- 1st Place Winner at ETHGlobal Delhi 2025 (Filecoin Track)
- Fully decentralized synthetic dataset creation and verification workflow
# What it is
FileThetic is a decentralized marketplace for generating, cryptographically verifying, and trading synthetic AI datasets. Built on Filecoin and IPFS, it enables model trainers to source verifiable synthetic data with provable provenance.
# Why it exists
Frontier AI models are depleting clean human-generated web data, while synthetic datasets lack provenance, quality verification, and tamper-evident licensing.
# How it works
Synthetic datasets are generated through quality-scored pipelines, persisted on Filecoin/IPFS, registered via smart contracts that anchor dataset proofs on-chain, and licensed through ERC-721/1155 provenance tokens.
# Details
- Filecoin and IPFS persistent decentralized dataset storage
- Smart contract dataset verification with on-chain cryptographic proofs
- ERC-721/1155 dataset license provenance tracking
- Synthetic data generation quality scoring pipelines
# Built with
Filecoin · IPFS · Solidity · Next.js · AI Pipelines