Skip to content
Preview — digitized from the official gazettes and checked automatically against the page each row cites. 831 of 832 FSSAI ingredient rows have been signed off by a person — some in bulk from the machine extraction rather than read page by page, so confirm against the gazette before you rely on a result. The 732 CDSCO rows behind a drug classification carry no human review at all.
Regulation bulletinLegal Metrology (Packaged Commodities) Rules — amendments changed since the last sweep — review for schedule or limit changes. (under review — not yet applied)23 Sept 2026

Government sources last swept 17h ago. 2 of 7 sources could not be fetched on the last sweep (FSSAI notifications (gazette, draft, ordinary), FSSAI nutraceutical regulation amendments), so changes to them would not have been detected. Verify against the gazette before relying on this. Detected changes are triaged by a human before any data moves.

Dolly Intelligence

Check a nutraceutical formulation

Nutra module · verdicts cite the exact schedule entry.

Delivery format · Regulation 4(1) — delivery format (source PDF p.6)
Or upload a formulation document

PDF, PNG, JPG, WEBP, GIF, TIFF or .txt — up to 50 MB with a live text layer, 20 MB if the words have to be read off the picture, which every TIFF is (converted to PNG first, since Anthropic’s API has no TIFF format of its own). The file is stored under your own account in Supabase Storage for the length of the read and deleted afterwards; if it has no live text layer, a picture of each page is also sent to Anthropic’s API to be transcribed. Nowhere else.

A dose read off a picture is marked amber, never green, until you edit the dose or unit yourself. A document’s own declared text (a real PDF text layer, or a .txt file) is read exactly, with no such marking.

Ingredients

Part and preparation are optional. Schedule II states a different allowance for each, so naming yours gets a determinate answer instead of the strictest reading.