Convert spreadsheets and documents locally

CSV, JSON, and TSV convert into XLSX or ODS. This is the step that usually involves opening a desktop application, waiting for it to start, importing with a dialog, and saving — and it is now a browser tab instead.

JSON is the interesting one. Turning arrays of objects into a table is exactly the operation people write throwaway scripts for, and having it as a validated conversion route removes a small recurring chore.

XLSX and ODS convert to each other with multi-sheet extraction, which matters because most real workbooks are not a single sheet. A converter that flattens a workbook to its first tab has quietly lost most of the file.

Check the sheet count after converting anything multi-tab. It is the fastest sanity check available and it catches the failure mode that costs the most time to discover later.

Spreadsheet conversion is not lossless in every direction. Formulas, conditional formatting, macros, and pivot tables are all things that either do not survive or do not survive faithfully between formats, and no converter can invent syntax the target format lacks.

Values and structure are the reliable part. If a workbook's value is in its formulas rather than its numbers, converting it is the wrong move regardless of the tool.

TXT, Markdown, and HTML convert to DOCX, PDF, or EPUB. Markdown to DOCX is the one that quietly saves the most time: writing in Markdown and delivering in Word is an extremely common workflow with an annoyingly manual middle step.

Markdown to EPUB is the path worth knowing about for anything long. An EPUB is a genuine ebook that reflows properly on a reader, which a PDF conspicuously does not.

Every route validates its output before enabling the download — real signature and container checks rather than a renamed file. On top of that, a build-time check refuses to publish a route that has no implemented, output-validating engine.

That second part is the unusual one. It means the conversion directory at /conversions lists only routes that genuinely work, so you should never meet a listed conversion that quietly fails.

The queue takes different file types at once and lets you choose an output per row, so a batch can send a CSV to XLSX, a Markdown file to DOCX, and a JSON file to a spreadsheet in a single pass. Each result downloads independently once it passes validation.

A failed row does not take the batch down with it — successful conversions stay available, which is the behaviour you want when one file in thirty is malformed.