SQL scratchpad
Query a CSV, Parquet or JSON file with real SQL. DuckDB runs inside this tab, so the file is never uploaded.
Your data
Files are read straight off your disk by the engine in this tab. They are not copied anywhere, and closing the page discards everything.
Your query
Result
What this is, and what it is not
This is DuckDB, not your warehouse. Its SQL is close to PostgreSQL, so Athena, T-SQL and Oracle specifics will not run here — use it to check that logic is right against sample rows, then take the query to the engine it belongs to.
Nothing is uploaded, and that is checkable rather than promised. The engine itself is served from this same origin, so even loading it makes no request to anyone else. Watch the network tab while you drop a file: it stays silent.
Query Parquet Files in Your Browser with DuckDB WASM
Apache Parquet is the industry standard for columnar data storage, but inspecting a file usually requires firing up Python, installing Pandas or DuckDB, or spinning up a Jupyter notebook. SQLParity brings the full power of DuckDB WebAssembly directly to your browser tab.
100% Client-Side
The browser File System API reads the Parquet file bytes directly into DuckDB's in-memory WebAssembly instance. No network requests are made.
Full Columnar Speed
DuckDB reads column metadata, statistics, and individual row groups on-demand, enabling fast aggregations, filters, and joins even on multi-hundred megabyte files.
Export Anytime
Run your query, inspect the results in the table, and export query subsets back to CSV or TSV with one click.
Security & Compliance Guarantee
Internal customer records, transaction logs, and analytical extracts should never be uploaded to random online file viewers. SQLParity enforces a strict browser Content Security Policy (connect-src 'self') to guarantee that your Parquet data never leaves your machine.