Importing inspection PDFs from other systems.

Section 01How to do it

1. 01

   Decide whether you need the import at all

   The cleaner route is to invite the inspector into your workspace: capture is native, the record is yours from the start, and you revoke access when the contract changes. Import is for the contractors who will not move.

2. 02

   Upload the PDF

   One report at a time. It is read for the property, the doors, the components, the pass/fail results and the notes.

3. 03

   Review before you commit

   This is the step people skip and regret. Check the property matched the right one, check door identities against your register, and check that a component read as failing actually failed.

4. 04

   Commit to the master record

   Once committed, the doors carry the imported history and behave like any other record: remedials, quotes, reports and exports all work from them.

5. 05

   Send anything odd to support

   Layouts vary wildly between vendors. If a report extracts badly, send it over — that is how the extraction gets better for the next one.

Section 02What the app actually does here

• Uploads are capped at 20MB per PDF
• Text-layer extraction runs first; the OCR fallback is beta, and reads at most six pages
• The parser reports a confidence score capped at 90% — it never claims certainty
• Property, address, inspection date, inspector, contact and requested-by are all extracted
• Imported components land on the same limit as captured ones: up to three failure reasons
• If no doors are detected you are told to add them by hand rather than left guessing
• An unrecognised layout can be sent to support as a template setup request

Section 03What goes wrong

Every one of these has cost somebody real time. They are cheap to avoid and expensive to unwind.

Section 04Questions

Section 05Related