← Latest products

Misata

synthetic data engine using LLM and Vectorized NumPy. Declarative synthetic data for Python. State the facts a dataset must hold (monthly revenue, channel shares, balanced ledgers, foreign keys, state machines) and get multi-table data that holds them exactly, with a DuckDB certificate anyone can rerun. From YAML, SQL DDL or a sentence. CLI and MCP server. I built Misata because existing tools (Faker, Mimesis) are great for random rows but terrible for relational or temporal integrity.

CLIOpen source

72 stars · GitHubexternal site

Visit github.comexternal site

Reactions (1)

Similar products

by description similarity

  1. DDL to Data

    news.ycombinator.com

    Generate realistic test data from SQL schemas

    Hacker News55 pts

  2. Xsql

    github.com

    Convert SQL Schemas Across MySQL, Postgres, and SQLite

    Hacker News11 pts

  3. I built an open-source data pipeline tool in Go

    github.com

    Build data pipelines with SQL and Python, ingest data from different sources, add quality checks, and build end-to-end flows.

    Hacker News200 pts

  4. New kind of data notebook for SQL

    hyperquery.ai

    Python, and viz

    Hacker News12 pts

  5. Curator

    github.com

    an open-source library for synthetic data generation

    Hacker News13 pts