Tender documents

GET /tenders/{tender_id}/documents lists the source documents the customer uploaded into the tender — the same attachments a forwarded LV / RFQ email carried (GAEB, PDF, Excel, images, .eml), or the files a POST /tenders upload sent. Use it to archive the customer's original request alongside the priced offer you write back into your ERP.

GET /tenders/63365/documents

{
  "data": [
    {
      "id": "b3f1c2a4-9d6e-4f21-8a0c-2e5d7f9a1b34",
      "filename": "LV_Tiefgarage_Beleuchtung.gaeb",
      "content_type": "application/xml",
      "size_bytes": 284913,
      "created_at": "2026-07-08T09:14:00Z",
      "download_url": "https://s3.fra1.example.com/mercura/files/…?X-Amz-Expires=3600&X-Amz-Signature=…",
      "download_url_expires_at": "2026-07-08T10:14:00Z"
    }
  ]
}

Only the customer's own uploads are listed. Mercura-generated artefacts — the offer / display PDF, the structured-doc preview, and internal OCR / metadata side-files — and any hidden files are excluded. If a tender has no uploaded documents, data is an empty array.
download_url is a short-lived, pre-signed link to the bytes in object storage. Fetch it with a plain GET — no Authorization header — and the file downloads with its original name and content type. The API never proxies the file contents.
The link expires at download_url_expires_at (one hour). Treat both the download_url and the whole document set as transient: re-list to obtain fresh links rather than persisting a URL. In non-production environments where the storage backend cannot sign URLs, download_url and its expiry are null.
id is the file's stable id: it is stable across re-lists, so you can skip a document you have already imported.
An unknown, non-tender, or foreign id returns 404 — indistinguishable from a missing tender.