Data handling
Public browser tools and authenticated Reconcile use different data paths
The boundary is intentional: a quick public utility should not upload your spreadsheet, while a saved reconciliation job requires explicit authenticated storage.
Public tools: files stay on your device
CSV, TSV, JSON and supported XLSX files selected on public tool and converter pages are read by JavaScript in the browser. The public tool form is not a multipart upload form and there is no public spreadsheet file-processing POST endpoint. Result CSV/XLSX/JSON files are built as browser downloads.
Optional public result links
Nothing is published automatically. If you choose Publish result, the browser sends only the summary, displayed headers and at most the first limited set of result rows. The source files are not transmitted with that action. Formula permalinks similarly store typed formula parameters and generated output rather than a workbook.
Authenticated Reconcile: deliberate uploads
When you sign in and start Reconcile, source files are intentionally uploaded because the product supports saved jobs, multi-source processing, server exports, audit records, profiles and scheduled summaries. Uploads are validated by extension, MIME, plan limits and worksheet structure and are staged in protected storage outside direct public browsing.
Protected exports and cleanup
Reconcile exports use private routes and noindex headers. Temporary workspaces and staged source files are eligible for configured cleanup. Audit records preserve operational evidence without turning the original files into public URLs.
User responsibility
Process only data you are authorised to use. Verify key selection, dates, identifiers, totals and business rules before using any comparison or export for payroll, payment, tax, HR, inventory or compliance decisions.