clmSpace
Documentation
Workflows

Ingestion

How a document in your Microsoft 365 SharePoint becomes structured, citable obligations you can act on.

The pipeline

  1. Bind a SharePoint folder under Settings, choosing whether it holds templates or agreements.
  2. Watch. clmSpace watches each bound folder and picks up new files automatically, reading them in place through Microsoft Graph. Your source files stay in your own SharePoint; clmSpace keeps no separate copy of them.
  3. Extract. Each file is parsed in place. PDF (with OCR for scanned pages), Word, Excel, CSV and text are all supported. Claude Sonnet then returns structured obligations with verbatim citations, following a fixed schema so the output is consistent.
  4. Persist. The derived structure is recorded as an agreement, its obligations and the verbatim citations (with the location on the page), each linked back to the source SharePoint file.
  5. Analyse. A second Claude Sonnet pass compares each obligation to the resolved standard position for its scope and flags any deviations for review.

What you see in the UI

Settings shows live per-file progress as each document moves from queued to ingesting to done. When a file needs attention, the outcome and the reason are recorded against it so you can resolve it directly.

Re-ingesting

Choose Re-ingest on a bound folder to replay every file in it. This is useful after you refine your standards or when you have moved a file from Templates to Agreements and want it re-read in its new role.

One way in: bound SharePoint folders
Binding a folder is all a new tenant needs. clmSpace reads the documents you already keep in Microsoft 365, in place, so there is nothing extra to deploy and your files never leave your own tenant.

Related