Review

Upload your file and ask us to review it. No AI: we return the same file, fixed (in its format, with its PDF) and a report with everything we found and everything we did. It works for Word (.docx), PowerPoint (.pptx), Excel (.xlsx) and PDF. It also takes the old .doc, .ppt and .xls formats: we convert them and the result comes back in the modern format (.docx, .pptx, .xlsx).

Upload the file

curl https://api.eu.doconda.com/v1/files \
  -H "Authorization: Bearer $DOCONDA_API_KEY" \
  -F "file=@contrato.docx"
# → { "id": "file_01JAB3…", "filename": "contrato.docx", … }

Ask for the review

POST /v1/documents
{ "operation": "review", "file": "file_01JAB3…" }

Like any document, in direct, stream or background mode. If Doconda made the file, pass its id instead of uploading it: "file": "doc_01JAB3…".

Read the result

status tells you how it turned out and GET /documents/{id}/report gives you the details.

`status`	What it means
`ready`	No problems (or all fixed).
`ready_with_warnings`	Minor warnings remain.
`needs_review`	Something remains that we can’t fix without changing your content: check it in the report.

We never change your text or your values: a mistyped amount is flagged, not corrected. We only apply safe fixes; the rest is reported. Each fix is applied only if the file improves; otherwise it’s discarded and noted in the report (ledger).

Where each problem is

Each issue in the report says where it is, in location:

pages: the pages (Word, PDF) or slides (PowerPoint), starting at 1.
sheet and cells: in Excel, the sheet and the cells or ranges (B3, A2:B2).

{
  "code": "xlsx.formula_error",
  "severity": "error",
  "message": "Sheet \"Sales\" has cells showing the error #REF!: B7.",
  "fixable": false,
  "location": { "sheet": "Sales", "cells": ["B7"] }
}

Security

Before opening your file we check it for security:

We reject files with macros, files that aren’t really the format they claim, password-protected PDFs, abnormal internal structures or XML declarations (common attack vectors).
We remove anything that would load or run by itself: linked images and links to external resources, attached templates, embedded objects and files, automatic actions on open and links that launch a program. Normal hyperlinks are kept.

Anything we remove appears in the report, in sanitized.removed.

What we check

In every format: money.words_mismatch (an amount in words that doesn’t match the figure, such as “€1,200” versus “one thousand three hundred euros”) and font.missing (a font we don’t have). Both are reported, not fixed.

Word (`.docx`, `.doc`)

Code	What it detects	Automatic fix
`table.overflow_width`	A table wider than the page	Fit to width
`heading.orphaned`	A heading alone at the bottom of a page	Keep it with the next paragraph
`signature.split_across_pages`	A signature split across pages	Keep it together
`numbering.discontinuous`	Gaps in clause or section numbering	Renumber
`style.not_applied`	The requested style doesn’t appear in the result	Reapply the style
`crossref.broken`	”see clause 5” that doesn’t exist	— (reported; and we don’t renumber, so it doesn’t point to another one)
`content.empty_section`	A heading with nothing under it	—

PowerPoint (`.pptx`, `.ppt`)

Code	What it detects	Automatic fix
`pptx.text_overflow`	Text overflowing its box	Shrink the text to fit
`pptx.shape_off_slide`	A shape partly off the slide	Bring it back inside
`pptx.empty_placeholder`	An empty placeholder	Remove it
`pptx.shape_outside_slide`	A shape entirely outside the slide	—
`pptx.text_too_small`	Text below 12 pt	—
`pptx.too_many_fonts`	Too many different fonts	—
`pptx.empty_slide`	An empty slide	—
`pptx.no_title`	A slide without a title	—
`pptx.duplicate_title`	Repeated titles	—
`pptx.image_low_resolution`	A picture below 96 dpi	—
`pptx.image_linked`	A picture linked to an external file	—
`pptx.no_slide_numbers`	A deck of 10 or more slides without numbers	—

Excel (`.xlsx`, `.xls`)

Code	What it detects	Automatic fix
`xlsx.number_as_text`	Numbers stored as text	Convert them, when unambiguous
`xlsx.print_too_wide`	A sheet too wide to print	Fit to page width
`xlsx.formula_error`	Formula errors (`#REF!`, `#DIV/0!`…)	—
`xlsx.broken_reference`	Broken references	—
`xlsx.external_link`	Formulas that depend on other files	—
`xlsx.inconsistent_formula`	A number typed by hand between matching formulas	—
`xlsx.date_as_text`	Dates stored as text	—
`xlsx.hidden_data`	Hidden sheets, rows or columns with data	—
`xlsx.merged_in_table`	Merged cells inside a table	—
`xlsx.empty_sheet`	An empty sheet	—
`xlsx.no_header`	A table without a header	—

PDF

Code	What it detects	Automatic fix
`pdf.no_text_layer`	Scanned pages without text	Add an invisible text layer (OCR) so it can be searched and copied
`pdf.title_missing`	No title	Set it from the first page
`pdf.language_missing`	No language	Set it, detected from the text
`pdf.fonts_not_embedded`	Fonts not embedded	—
`pdf.blank_page`	Blank pages	—
`pdf.page_size_mixed`	Pages of different sizes	—
`pdf.text_rotated`	Text sideways or upside down	—
`pdf.image_low_resolution`	A picture below 150 dpi	—
`pdf.untagged`	An untagged PDF (accessibility)	—
`pdf.link_broken`	Broken internal links	—
`pdf.form_empty`	Empty form fields	—
`pdf.annotations`	Comments or review marks left in the document	—

Get started

Operations

Concepts

Account

Where each problem is

Security

What we check

Word (`.docx`, `.doc`)

PowerPoint (`.pptx`, `.ppt`)

Excel (`.xlsx`, `.xls`)

PDF

​Where each problem is

​Security

​What we check

​Word (.docx, .doc)

​PowerPoint (.pptx, .ppt)

​Excel (.xlsx, .xls)

​PDF

Where each problem is

Security

What we check

Word (`.docx`, `.doc`)

PowerPoint (`.pptx`, `.ppt`)

Excel (`.xlsx`, `.xls`)

PDF