{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/14307b529e3a4b829c1d4daf9fbd1aed\" frameborder=\"0\" width=\"1398\" height=\"1048\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1048,"width":1398,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1048,"thumbnail_width":1398,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/14307b529e3a4b829c1d4daf9fbd1aed-e7ef3694fd419e99.gif","duration":136.738,"title":"Parakh GST Credit Checks","description":"Parakh is an evidence-first counterparty-checking experience for Indian small\nbusinesses. Before sending goods on credit or starting a business relationship,\nan owner may need to compare GST filing patterns, name variants, public-record\nclues, screenshots, and portal results. Today, that work is fragmented, slow to\nread, and easy to overstate.\n\nThis public, no-login prototype makes one check easier to complete. A reviewer\nenters one visible Demo Reference, watches the local-fixture report flow, and\nreceives a readable Parakh-style report. It turns fictional registration,\nfiling, and public-record fixtures into explainable FLAG, CLEAR, and NOTE\nobservations. Each observation shows confidence, attribution, provenance,\nlimitations, and “What we could not find.” The user can compare five realistic\nfictional situations, inspect the evidence path, or download a fixture-only A4\nPDF.\n\nThe difference is not a score or a trust verdict. Parakh makes the available\nevidence and its gaps understandable before a business owner acts. It is built\nfor mobile use and lower digital confidence: one clear reference, one visible\njourney, and plain-language limits.\n\nEverything in this build is synthetic. It never accesses live government\nsystems, private records, real identifiers, payments, OTPs, or production\nParakh data. A real deployment would require authorized APIs, consent-aware\nhandling, audit logs, rate limits, data provenance, retention limits, security\ncontrols, and clear human-readable limitations."}