Query CSV and Parquet files locally with SQL and save a reusable analytical database.
Free MIT engine · own CPU, RAM and disk · no hosted compute included

Developer tools
Run Python analysis in saved notebooks with outputs you can reopen and export.
Run Python analysis in saved notebooks with outputs you can reopen and export.
Replace individual Python notebook analysis with local kernels and saved outputs. No managed distributed compute, Unity Catalog, production jobs, shared governance or hosted service. Own hardware and any hosting or APIs are extra.
Local notebook analysis only: install and maintain Python, kernels and dependencies. Not a distributed Spark service, governed shared warehouse, managed scheduler or production model-serving replacement. Your hardware, electricity, backups and any hosted deployment cost extra. Cloud/API calls made by your code are separately billed; none are needed for local analysis. Verification covered headless execution and saved outputs, not the browser editor, extensions, remote kernels or collaboration. Do not expose a notebook server publicly without authentication and secure transport.
Each comparison covers the workflow described above, rather than every paid feature.