Built and shipped.

Native iOS (SwiftUI) and Android (Kotlin, Compose) apps and a Next.js web app on one Supabase backend, so a profile, its verification state and its matches are the same on every device. Members verify their identity with a selfie before they appear as verified, and the badge attests to the primary photo: when the file behind a photo changes, it is screened again, not only on upload.

The marketing site ships an English and a Persian edition, one indexable URL per language with reciprocal hreflang, generated from a single source so the two never drift.

Verify it yourselfOpen the store listings, load lovejan.app and lovejan.app/fa/, then run the scan above.

A bilingual English and Farsi SwiftUI app with an adaptive iPad layout, on Supabase with row-level security, Stripe for payments and in-app calls. It operates as a California Domestic Referral Agency: families browse, choose and hire; the app matches and connects. Matching for IHSS families is free, as the law requires.

A bilingual resource library with structured key facts and sourced legal disclaimers on every article that touches money, law or health.

Verify it yourselfLoad carejan.com, read an article in either language, then run the scan above.

Next.js on Vercel. Every page is server-rendered with one JSON-LD graph; sitemap dates come from a page registry, never from the clock; robots.txt welcomes every answer crawler by name and repeats the API disallow inside each group, because a crawler obeys only its most specific group.

CI scans every rendered page for the kinds of claim that are easiest to invent (percentages, client counts, ranks, ratings, credentials, what clients say, and who an AI engine named) and fails when one is not tied to a current evidence file, or when a price is not in the published price list. NeshatScan itself is an SSRF-guarded fetcher that pins the resolved address at connect time, an RFC 9309 robots.txt parser that reports the matching line, and a set of checks streamed to the page as each one finishes.

Verify it yourselfView the source of any page, read robots.txt and llms.txt, then scan neshat.ai with its own tool.