{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreih7s2nsfnm32t627i5afrybt4w53lchboiuzfmqjvafqmsr2ntnda",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mp3lymcnh2y2"
  },
  "path": "/t/seeking-indic-document-dataset-india-invoices-receipts-utility-bills-payment-advices-packing-lists-commercial-invoices-credit-notes/177055#post_6",
  "publishedAt": "2026-06-25T03:04:39.000Z",
  "site": "https://discuss.huggingface.co",
  "tags": [
    "AgamiAI",
    "Voxel51",
    "shubh303",
    "mychen76",
    "AgamiAI’s Indian financial document datasets",
    "InvoiScope",
    "gstextract-core",
    "invoice2data"
  ],
  "textContent": "For now:\n\n* * *\n\nA few additional leads that may be worth checking:\n\n  * AgamiAI\n\n    * They seem close to the synthetic Indian financial-document space.\n    * Their PDF + JSON direction may be relevant here.\n  * Voxel51\n\n    * Not India/Indic-specific, but their invoice OCR / dataset QA work may be useful as a reference.\n  * shubh303\n\n    * Has invoice-to-JSON related dataset work.\n    * Probably more useful as a format/schema reference than as a direct match.\n  * mychen76\n\n    * Has invoice/receipt OCR-style datasets.\n    * Again, likely useful for structure and annotation-format ideas rather than Indic coverage.\n  * AgamiAI’s Indian financial document datasets\n\n    * Worth checking separately from the org profile, since their dataset cards may give clues about future or custom synthetic document generation.\n\n\n\nA few HF-adjacent / outside-HF leads:\n\n  * InvoiScope\n\n    * Indian GST invoice extraction project.\n    * Could be useful for field schemas, bbox-style annotation ideas, or referrals.\n  * gstextract-core\n\n    * Indian GST invoice PDF extraction project.\n    * More PDF/schema-oriented than scanned-image/bbox-oriented, but still close in domain.\n  * invoice2data\n\n    * Not India-specific, but a mature invoice extraction project.\n    * May be useful for template-based extraction ideas or weak-label generation.\n\n\n\nOverall, I still wouldn’t expect one public dataset to match everything directly. But these may be useful people/projects to check for synthetic data, schema references, or vendor referrals.",
  "title": "[SEEKING] Indic Document Dataset (India) — Invoices, Receipts, Utility Bills, Payment Advices, Packing Lists, Commercial Invoices, Credit Notes"
}