Reports and exports

Reports turn your extracted document data into rows you can read, filter, export, or hand to your accounting system. There are two kinds: the built-in reports — fixed views over your documents and your credit spend — and the report builder, where you assemble your own report by dragging in exactly the fields you want. All of it lives behind the Reports & exports tab strip, which groups pages under three headings: Operations, Cost, and Data. The sample invoice on your onboarding checklist never appears in any report, on screen or exported — it's a demo document, not real data.

Built-in reports

Summary

Summary (Operations tab) gives you one row per document — document number, counterparty, date, total, status — filterable by document type, counterparty, validation state, and date range. Use it to scan a batch at a glance, or to jump into a specific document from the row.

Its Lowest field confidence column shows the same weakest-field figure described under The document view — not the model's overall self-reported score — coloured green at 85% and up, amber at 65% and up, and red below that.

Line Items

Line Items (Operations tab) is the same filter set, but one row per line item instead of per document — description, quantity, unit price, line total, VAT rate. Use it when you need to check or export the detail under the totals, not just the totals themselves. It covers every built-in document type, plus custom-template documents — pick Custom (all templates) or a named template from the document-type filter the same way you'd pick any built-in type.

Statements are the one type that works a little differently. A vendor or customer statement has no line items — it has a ledger of transactions, each one referencing an invoice, payment, or adjustment that lives elsewhere. Line Items gives you one row per ledger entry, in its own set of columns: date, description, type, reference, the referenced document number, debit, credit, signed amount, and running balance. They stay separate from the invoice line columns on purpose, so that adding up a column of line totals never accidentally counts an invoice twice — once as the invoice itself, and once as the statement row that mentions it.

For a custom-template document, the on-screen preview still shows only the fixed set of columns above (it's a shared table shape); your template's own fields — everything you defined in the field designer — appear in the downloaded Excel or CSV, not the preview. Download the report if you need to see the full column set.

Both Summary and Line Items support the same filters (document type, counterparty, date range, validation outcome) and both export from the Export to Excel and CSV buttons at the top of the page — the export runs over every matching row, not just the ones on screen.

Bank account numbers and IBANs show only their last four digits (for example ••••7890) in both reports, on screen and in the downloads, for a Reviewer unless an admin has granted them bank number access; Admins and Firm admins see full numbers. The same goes for a "New bank account" check quoted in the validation column.

Credits

Credits (Cost tab, requires the Admin role) shows your current credit balance and month-to-date consumption, broken down per document — including which operations consumed credits on that document (an extraction, a retry, a fix). Where a document failed validation for a reason no further extraction attempt would resolve — for example an unrecognised counterparty — the report says so directly, so you know to resolve it by hand (approve or create the counterparty) rather than waiting on a retry.

If your workspace is part of a firm's pooled billing, the credits shown here are still scoped to this workspace — see Billing and credits for how the shared pool works across a firm's clients.

Models and Benchmarks

The Operations tab also lists Models and Benchmarks. These are platform-administration pages for comparing extraction-model performance across all workspaces on DocReadi — they are not part of a normal workspace's day-to-day reporting and most workspace administrators will never see them.

Per-document export

From the document viewer, or from a selection of documents on a list page, the Download button opens a chooser with four options:

• PDF file — the original document (a .zip when you've selected more than one).
• Summary report — the summary view described above, scoped to your selection.
• Line-items report — the line-item detail, scoped to your selection.
• Markdown — the parsed text DocReadi extracted from (a .zip for multiple documents). For a reviewer without bank number access, the bank account numbers and IBANs in that text show as ••••7890, the last four only.

For the two report options, a second choice appears: Excel or CSV. Pick a kind, pick a format if relevant, and select Download. The report downloads treat bank account numbers the same way as the reports above: masked for a Reviewer unless granted access, full for Admins and Firm admins.

The report builder

Reports & exports (the Data tab, and also the page you land on from the Reports & exports link in the main navigation) is where you build your own report. It requires the Admin or Reviewer role to view; only an Admin can create, edit, or delete a saved report.

Built-in exports

At the top of the page, DocReadi keeps a small set of pre-built exports tuned to a specific destination format rather than to a workspace's own columns — for example Xero Bills (AP invoices), which downloads the AP invoices of your approved documents as a CSV in Xero's Bills import format (one row per line item, vendor/invoice number/dates/amounts pre-mapped, and the GL account code filled in where GL determination has assigned one). Credit notes and sample documents are left out. An invoice is picked up by its invoice date; one with no invoice date is picked up by the day the document was received and exported with an empty InvoiceDate column, so Xero's import will ask you for a date. When an invoice has no due date, the DueDate column uses the invoice's own printed payment terms (worked out with their basis — Net 30, month end, "30 days from statement", "Net 30 EOM" and so on), or — when it prints no terms — the supplier's own payment terms, and after that your workspace's default payment terms. Printed terms that cannot be read leave the column empty rather than being replaced by an assumption. See Default payment terms for how that's worked out. These sit above your custom reports because, for many workspaces, they're the export used every day. Select Download CSV to run one immediately; it defaults to the last 90 days.

Building a report

Select + New Report to open the builder. The builder has three parts:

• Available Fields (left) — every column you can add, grouped into Most Used, Smart Fields, Document Fields, Extracted Fields, Metadata, Counterparty, Review requirements (if your workspace has any active — see below), and — if your workspace has custom templates — one group per custom template's own fields. Use Search fields… to find one quickly. The Metadata group's confidence field is listed as lowest_field_confidence — the same weakest-field figure as the Summary report's column above — and exports as a plain 0–1 number, blank when a document has no confidence. The Summary and Line Items downloads carry the same figure in their confidence_score column.
• Columns (right) — the fields you've added, in the order they'll appear. Drag a field from the available list to add it; drag within the column tray to reorder; remove a column you no longer want. Beyond the fields on the left, you can also add an Add Calculated Column (a formula over your other columns, e.g. doc_total * 0.15, validated as you type) or an Add Constant Column (a fixed value repeated on every row).
• Filters — Document type, Status, Source, Folders, Date from, and Date to, applied before the preview and the export run. Source scopes the report to documents that arrived from one origin — pick a specific inbox source by name, or a channel such as Web upload, WhatsApp, Email-in, or Manual / API — or leave it on All sources. Folders scopes the report to any number of your own filing folders — pick as many as you need from the checkbox dropdown; each chosen folder covers itself plus everything filed beneath it, and a document in any chosen folder is included. The closed dropdown reads All folders with none picked, the folder's name with one picked, or "N folders" with several; a folder you've since emptied still shows here if a saved report still points at it. A Unique only checkbox suppresses duplicates, reissue losers, and re-uploads so you see one row per canonical document.

Give the report a name in the Report name… field (and, optionally, a description), then select Save Report. Use Refresh at any point to preview the current column and filter selection before you save.

Bank account numbers are masked to the last four digits in report previews and exports for a Reviewer who hasn't been granted bank number access — the same masking rule the document viewer applies, including where a text column (an address, a line description, a validation message) quotes the account or an IBAN. Admins and Firm admins see full numbers.

Review requirement fields

If your workspace has review requirements configured, the Review requirements group lists one field per active requirement, named after its label. An Input — Number requirement exports the coerced numeric value rather than the reviewer's raw text; an Input — Yes / no question exports the reviewer's answer as the text yes or no; a Confirm requirement offers three fields instead of one: "label (confirmed)" (true while a live, non-invalidated confirmation exists), "label (confirmed at)", and "label (confirmed by)" (the reviewer's email). If the field being confirmed changes after the tick — the panel's changed since confirmation state — the confirmation exports as blank or false rather than as the value it once attested to. Deactivating a requirement removes it from the field picker for new reports, but a saved report that already references it keeps working.

A requirement whose Applies to covers several document types still shows as a single entry here, not one per type — its column fills from whichever document type each row in your report actually is.

Running and exporting a saved report

From the report list, each saved report offers:

• Run — opens the report's rows in the browser.
• CSV / Excel — downloads the report in that format directly.
• Edit — reopens the builder with the report's current columns and filters (Admin only).
• Delete — removes the saved report (Admin only).

The Run page repeats the same filter bar (Document type, Status, Source, Folders, Date from, Date to, Unique only) above the rows, so you can look at this run through a narrower or wider lens without editing the saved report itself. The Folders dropdown pre-checks the report's saved folders; ticking or unticking a folder changes this run only, and unticking every folder means all folders for this run. Leave a filter as it is to keep the report's saved value; set one to override it for this run only; Reset clears every override and returns to the report exactly as saved. Choosing Edit now opens the builder with every saved filter — including Source and Folders — already selected, matching what the report actually runs against.

Once your overrides make this run's filters differ from what's saved, a bar appears above the rows: "This run's filters differ from the saved report." Enter a name (optional) and select Save as new report to keep this exact scope as its own saved report, with the same columns as the original — the original report and its filters are never changed. Leave the name blank and DocReadi builds one for you from the original report's name and a short label for the scope, such as the folder or source you narrowed to. Saving takes you straight to the new report's own Run page. This needs the Admin role, the same as creating any report.

The reports API

Everything a saved report can do — list your reports, discover the available fields, create, update, delete, run, and export as CSV — is also available over the REST API using an API key, so you can automate a report pull from your own systems rather than exporting by hand each time. See /api/docs-guide for the full endpoint reference and authentication details.

Next: Delivering documents — get your approved data (and these reports) into the systems downstream of DocReadi.

Looking for the developer/API reference? See /api/docs-guide. Can't find an answer? Email [email protected].