Sign in to continue

Extraction is limited to authorised accounts. Files you upload are private to your account and are removed once your download link expires.

Password

Large documents can be run without an account using the command-line extractor.

Turn a document into structured JSON

Headings, hierarchy and body text are segmented into addressable sections for RAG ingestion or compliance automation. Nothing is retained after your download link expires.

Extraction runs while you wait and is capped at 29 seconds. Documents over roughly 100 pages are better run through the command-line extractor, which streams and has no time limit.

PDF

Uploaded as

Storage accepts letters, digits, dot, dash and underscore. Your document is unchanged — only the stored name differs.

Document ID Defaults to the filename. Letters, digits, dot, dash and underscore only.

Profile Chooses the heading grammar used to find sections. Commercial policies, vendor attestations and ISO or SOC 2 documents go through Generic — the other three expect a government numbering scheme.

Rows become

One section per row — default One section per sheet Chunks of rows

Per-row suits control matrices; per-chunk keeps large exports under the cap of 250,000 rows.

Group orphaned content under virtual parents When a document references a section it never defines, create a placeholder so child sections aren't orphaned. Include table-of-contents entries Off by default — TOC lines usually duplicate the headings they point at.

Runs while you wait, under a 29-second service ceiling.

0% Uploading to secure storage…

upload link valid for 300 s · direct browser → S3

Segmenting sections…

0:00

Large documents can take longer than the service will wait. Nothing is lost if it does — you can re-run or use the command-line extractor.

Still processing after 29 seconds

The service answered before your document finished. Extraction may still be completing in the background, but this page has no way to collect the result — there is no job-status endpoint yet.

Documents this size are better handled by the command-line extractor, which streams and has no time limit.

How to run it locally

link expires 15 minutes after extraction

Only one section found. This document may not use heading styles the selected profile recognises — try the Generic profile.

Sections