Core Concepts

  • Dataset — A named collection you upload (tabular CSV/Parquet, JSON/JSONL, time‑series, images).

  • Synthesis Job — Generates a synthetic copy with privacy & quality options.

  • Quality Report — Fidelity diagnostics (distributional, relationships, task parity).

  • Privacy Report — Risk diagnostics (NNDR proxy, k‑anonymity, nearest‑neighbor leakage).

  • Agents — Autonomous components that infer schema, detect edge cases, choose generators, and validate outputs.

  • Data Hub — Contribute and license datasets with on‑chain provenance and royalties.

Last updated