Turn unfamiliar files into understood, useful data.
Import structured data, inspect its real quality, reshape it visually or with SQL, compare multiple datasets, and export the result. DuckDB-Wasm runs the work in a dedicated browser worker.
Selecting the best WebAssembly bundle and opening local storage.
CSV, TSV, JSON, NDJSON, XML, YAML, XLSX, Parquet, Arrow and SQLite. Everything is processed locally in your browser — files never leave your device.
- 01
Drop file
Ten structured-data formats
- 02
Understand data
Schema, types, nulls, duplicates and malformed rows
- 03
Transform or query
Visual recipes or local DuckDB SQL
- 04
Preview result
Paginated, typed result inspection
- 05
Export
Ten local output formats
Import broadly
CSV, TSV, JSON, NDJSON, XML, YAML, XLSX, Parquet, Arrow and SQLite.
Shape precisely
Filter, sort, rename, cast, deduplicate, clean nulls, flatten, pivot and calculate.
Work across files
Join, union, split, compare schemas and generate a row-level dataset diff.
Local-first answers
How does Data Studio handle a file?
The browser gives the dedicated worker access to the selected file. DuckDB-Wasm and format-specific parsers inspect it locally, derive a schema, and calculate quality signals without uploading source rows. When OPFS is available, the workspace persists on this device; otherwise the interface clearly reports a memory-only session.
What appears in the instant profile?
Every import immediately reports row and column counts, inferred types, missing values, exact duplicates, minimum and maximum values, frequent values, histograms, date ranges, encoding and file statistics. Potential malformed data and deterministic fixes appear beside the evidence, so cleanup decisions remain reviewable rather than automatic or hidden.
When does data leave the browser?
Source files never leave the browser. Downloads are generated locally. Only an operator who selects a compatible PlatPhorm destination, provides the shared platform key, and presses Publish transfers the selected generated export. The protected relay uses fixed destinations, enforces format and size limits, and reports success only after the destination returns durable evidence.