File formats and limits

Source inputs

Sources accept the following inputs.

Document file

PDF, DOCX, TXT, or Markdown. Files contain at least one byte and at most 10 MiB. A PDF contains at most 100 pages.

Pasted text

Text entered in the source form. The input field accepts up to 500,000 characters. The saved source also has the 10 MiB limit.

Owner attestation

A firsthand statement entered by the account owner. The source retains its attestation label.

Source URL

Optional origin metadata. River does not fetch the URL.

Image ingestion and optical character recognition, or OCR, are not supported. Image-only scans do not provide selectable text for extraction.

An imported résumé becomes a source. The import does not create a structured résumé draft.

Export files

An exported checkpoint retains these files.

PDF

The formatted résumé.

LaTeX source

The generated document source in a .tex file, available in Advanced tools.

Extracted text

The text recovered from the rendered PDF in a .txt file.

Validation report

The recorded document checks in a .json file, available in Advanced tools.

DOCX export is not available. DOCX import is supported.

Downloads require access to the account that owns the files. Public documentation does not grant access to workspace downloads.

Originals and historical files

River preserves original source content and provenance. Each completed extraction has a separate identity. An extraction retry does not replace text in existing citations.

Draft preview files can expire. Checkpoint files and exported files remain available as historical outputs.