Skip to main content
Every processed document opens at /document/attachment/<id>/. This is the only document page — older /document/<id>/ and batch-document URLs redirect here.

​
Layout

The left pane is the original file (PDF viewer with bounding boxes, or the image). The right pane is a tab strip: Header chips show the filename, size, and applied schema. The toolbar toggles the PDF, bounding boxes, and type size.

​
Actions

  • Apply schema / Reprocess — pin a published schema (or clear it) and run extraction again. A completed document is extracted fresh; a failed retry resumes where it left off.
  • Create schema — jump to the schema builder with this file as context.
  • Share — public link.
  • Download — the original file.
  • Delete — remove it from your library.
  • Retry — shown when extraction failed.
On a public /s/<token>/ link the page is read-only: Sections and Schema only, no chat input, no edits, no share.
Schema fields and transcription are independent. You can read the text without applying a schema. Applying one and reprocessing fills the Schema tab.