FilAI

FilAI is a decentralized platform that automates dataset quality analysis, bias and anomaly detection, and cryptographically verifiable storage/verification of datasets and analysis results using Filecoin onchain cloud/IPFS and smart-contract-based registries. It let researcher (crypto or Academic) and Journalist verify their dataset and also monetize their dataset on the marketplace.

FilAI is built as a modular, decentralized data-analysis pipeline designed to combine traditional data-science tooling with verifiable on-chain proofs. The frontend is a React-based interface that handles dataset uploads, job orchestration, and visualization rendering. The backend is composed of lightweight Python microservices using FastAPI for ingestion, Pandas, Numpy for statistical profiling, and scikit-learn–based modules for anomaly detection, correlation analysis, and bias assessment.

For storage and verification, we integrated directly with Filecoin Onchain Cloud (FOC) and IPFS. Datasets are content-addressed, and we generate SHA-256 digests and Merkle proofs so the analysis results can be cryptographically tied to the exact dataset version. Filecoin’s Warm Storage and Beam retrieval layers dramatically improved performance allowing fast, verifiable reads without re-downloading entire datasets. We use FOC’s programmable payments to simulate recurring storage payments for long-lived data. A lightweight Solidity smart contract acts as a dataset registry where we record ownership metadata, IPFS/Filecoin CIDs, and hashes of analysis reports so anyone can independently verify integrity.

The SDK is written in Python and wraps the analysis and storage calls into simple functions, making it easy to script dataset uploads and quality scoring. One of the hackier parts was implementing a streaming chunk-based profiler so huge datasets could be analyzed without loading everything into RAM; another was using Merkle trees so we could anchor proofs on-chain without exposing raw data.

Overall, FilAI is stitched together from modern data tooling, decentralized storage primitives, and cryptographic integrity checks to create a transparent, verifiable dataset-analysis platform.