Document downloads
Download original registry documents or request a PDF rendition of HTML or a PDF archive.
Document orders return a downloadLink. Follow that link to download the original
registry file. The link is a bearer capability: anyone with it can download the
document without an API key. Keep it private to the intended recipients.
Request a single PDF
Append ?convertPdf=true to the downloadLink to request a single PDF, including
for orders placed before this option was introduced:
curl --fail --location "${DOWNLOAD_LINK}?convertPdf=true" --output document.pdfIf the URL already has a query string, append &convertPdf=true instead. The option
is available in every supported API version, with or without Kausate-Version.
| Source | With convertPdf=true |
|---|---|
| Ordinary PDF | Returns the original PDF bytes with application/pdf. |
| HTML or XHTML original | Renders the document as one PDF, retaining tables, text, author styles and page breaks. Scripts are not executed. |
| ZIP containing one or more PDFs | Returns one PDF containing every page, in ZIP entry order and original page order. Nested directory paths are supported; directory entries add no pages. |
| ZIP containing JSON, XBRL, XML, XHTML, HTML, images (including TIFF), office documents, signature files, or nested archives | Returns HTTP 422. These members are neither converted nor omitted. |
| Other source, empty archive, unreadable archive/PDF, or encrypted archive/PDF | Returns HTTP 422. No partial PDF is returned. |
The response uses HTTP 200, Content-Type: application/pdf, and an attachment
filename ending in .pdf. PDF recognition uses file contents rather than filename
extensions. The complete source must fit within 50 MiB, with at most 256 ZIP
entries, 100 MiB of expanded content, 100 MiB of output and 5,000 pages. Exceeding
a size or page limit returns HTTP 422. Conversion runs with a 256 MiB processing
memory budget on the deployed API; a document that exhausts this budget also
returns HTTP 422. A missing stored source returns HTTP 404.
Other storage or processing failures return
HTTP 500. Retry without the option to retrieve the original.
Merging copies PDF pages, including their text and images. Document-level digital signatures, bookmarks and embedded attachments are not preserved in the merged PDF. Use the original files when those features are required.
PDFs with repairable cross-reference offsets are accepted when their pages remain readable, including scanned Companies House accounts. A standalone PDF is returned byte-for-byte unchanged; repaired PDFs inside ZIPs contribute their pages to the joined PDF. An unreadable PDF still returns HTTP 422.
Omitting convertPdf, or setting it to false, retains the original HTTP 302
redirect and source file unchanged. The option does not modify the stored source,
order metadata, contentType, or fileName returned by the order. It does not
place another registry order.
HTML renditions use print styles and an A4 page when the document does not specify a page size. Pages widen or grow when fixed-size content would otherwise be clipped, up to 4,096 CSS pixels per dimension; layouts that still cannot fit, including CSS-transformed content, return HTTP 422. UTF-8 and other declared character encodings are preserved. Referenced stylesheets, fonts and images must be available over public HTTP/HTTPS or embedded as data URLs. Local files and private network resources are refused. Each resource is limited to 5 MiB, with at most 128 resources and 100 MiB total resource content; an unavailable or refused resource returns HTTP 422. Use the original HTML when interactive or script-generated content is required.
Jurisdictions and archive contents
The option applies to stored documents from every jurisdiction. Whether an archive can be merged depends on its contents:
- Netherlands KVK annual accounts can arrive as ZIPs containing PDFs.
- German Unternehmensregister annual accounts and Bosnian registry pages can arrive as HTML originals and render directly to PDF.
- Italy InfoCamere deferred deliveries can contain PDFs or TIFF scans alongside XML and text indexes. Packages with any non-PDF member return HTTP 422.
- Poland eKRS filings can contain several PDF attachments or mixed native files.
- Denmark CVR announcement downloads wrap the published PDF files in a ZIP.
- Mexico BMV annual reports arrive as ZIPs containing structured JSON.
- Sweden Bolagsverket, Estonia e-Business Register, Slovenia AJPES, Spain's commercial registries, South Korea OpenDART and Brazil CVM can deliver native archive packages.
- Hungary e-beszamolo returns ZIPs for mixed attachments or an unreadable PDF set; company-register evidence bundles can also contain non-PDF files.
- German financial disclosures and transparency-register evidence, and Minnesota business-record evidence, can include structured files, HTML or signatures.
For mixed packages, HTTP 422 means the original archive is needed to retain all files. Estonian BDOC/ASiC signature containers and ESEF packages are not converted to PDF. ZIPs used internally for bulk ingestion, or already unpacked/merged before document delivery (for example Slovakia), require no change to the download link.
Last updated on