Indian OCR Dataset Needed

I’m lining up an opportunity that hinges on finding a high-quality, off-the-shelf OCR dataset covering multiple Indian languages and scripts. If you already curate such data, I’d like to understand exactly what you have and see a representative sample this week so I can move the deal forward. Scope I’m interested in • Languages/Scripts: a broad mix—Hindi (Devanagari) plus regional scripts such as Tamil, Telugu, Kannada, Malayalam and any others you hold. • Document Types: invoices, receipts, business forms, and also books, handwritten pages or other scanned records. • Annotation Depth: full text, document layout (bounding boxes, reading order) and key-value pairs—for example invoice number, date and total. What I need from you now 1. A brief catalogue describing for each subset: language/script, document type, image/page count and the exact annotation layers included. 2. A zipped sample of 50 images with accompanying annotation files that demonstrate text, layout and key-value extraction. Please ensure the sample reflects the mix above. 3. Confirmation that rights are cleared for third-party commercial use. Timing Send the catalogue and sample package by end of this week; I’ll review immediately and revert with next steps. If your dataset meets the brief, this could turn into a larger licensing arrangement, so I’m keen to see what you can share.