Workflow

From one PDF to reviewed spreadsheet tabs

Choose the file once, preview the extracted tables, and decide what gets inserted.

1. Choose one local PDF

Open PDF to Sheets from the current spreadsheet and select one PDF from your device. The add-on does not browse arbitrary Google Drive files.

The file must be no larger than 50 MB and must fit your plan’s per-PDF page limit. The service never processes more than 100 pages or 250,000 extracted cells in one job.

Password-protected or encrypted PDFs fail safely. Do not send a password; the product does not accept or store one.

2. Conversion starts

There is no classifier or mode choice. Every supported file uses the same general-table extraction contract, including statements.

There is no page-range picker, batch upload, photo upload, mapping editor, or hidden normalization step.

3. Processing continues safely

The browser uploads directly to private temporary storage. Apps Script receives metadata, not PDF bytes. Jobs are durable and asynchronous, so closing the modal does not cancel work; the add-on can reopen a safe resume pointer while it remains available.

Extraction may be incomplete or wrong. Latency varies with document length, scan quality, and provider availability.

4. Preview before insertion

The result preview is read-only. It shows extracted tables, page provenance, and implemented fixed warnings—not a model confidence score. You select the tables to insert.

Source strings remain literal spreadsheet text. The product does not infer missing values, normalize dates or signs, reconcile balances, categorize transactions, or stitch unrelated rows.

5. Insert without overwriting

Selected tables default to separate new app-owned tabs. One small table can instead go into the current selection only when the entire destination is empty, unmerged, editable, and free of notes and validation.

Non-empty cells are never overwritten. There is no general one-click undo, so inspect the preview and destination first.

6. Temporary content is deleted

Codeprint-controlled PDF uploads are deleted as soon as reliable processing no longer needs them, and structured results after confirmed insertion. Abandoned content has a one-hour maximum availability target under the cleanup model.

Content-free account, usage, billing, security, and audit metadata lasts longer where operationally or legally required. Google’s paid Gemini service has separate limited abuse-monitoring retention described in the Privacy Policy.

Google Sheets add-on

Bring the PDF. Keep the final say.

Preview what was extracted before anything is inserted. Review every result—especially financial data.