Data WorkBench

One intake. Many sources.

Core model

Workspace -> Source -> Format Adapter -> Views -> Jig -> Conversion -> Receipt

Any-to-any means interpretation through a model and a ruleset, not implicit conversion. Export downloads are direct; optional details stay available for review.

  • 100% your controlOpen, edit, export, or save back only when you choose.
  • Complete code transparencyThis public workbench is static repo code you can inspect.
  • No harvestingThe code runs on your system. The data stays in your space.
Source trust cautions

Use known and trusted sources, but make trusted mean evidence: official domains, direct navigation, a known owner, current docs, and clear data handling. "I used it" and "they are big" are not enough by themselves.

  • USA guidance from FTC and CISA warns that familiar names and trusted-looking messages can still be impersonation.
  • UK, Australian, and Canadian authorities give the same practical rule: verify through official websites or contact details you found yourself.
  • When a source is unclear, inspect locally first and avoid pasting sensitive data into mystery tools.
  1. 1. IntakeUpload, paste, or load samples into separate source tabs.
  2. 2. InterpretDetect or choose JSON/JSONL, XML/WSDL, delimited, XLS/XLSX, code, or text.
  3. 3. ShapeUse Jigs for header rows, mappings, validation rules, and loss expectations.
  4. 4. ExportDownload or copy target output. Review details stay available when needed.

Load files

Upload, drop, or select sources

Unsupported extensions are hidden or blocked. Folders still open for browsing.

Active source

No source loaded

Load a sample source or paste source text to begin.