Entyti Medical · Pharmacy ERP
What each onboarded pharmacy costs to run, and where the shared infrastructure needs its next paid step. Built to answer one question: which costs get billed to the client, and which Entyti amortizes.
Scales directly with each shop's usage. This is the number Suheab's model recovers from the client.
| Driver | Unit rate | Source | Cost at busy volume |
|---|---|---|---|
| WhatsApp bill + PDF / reminder1 message per bill, invoice PDF attached | 14.5 p / msg incl GST |
ConfirmedAiSensy invoice: 13 msgs = ₹1.885 | ₹435 / mo100 bills/day → 3,000/mo |
| AI assistant (Haiku 4.5)quota-gated per shop | $1 in / $5 out per 1M tokens |
Anthropic pricing~₹0.35 / run est. | ₹0–525 / mosoft — not logged |
| PDF bill storage (Supabase)~8 KB/bill, 900s signed URL | ~8 KB / bill | MeasuredINV-97 → 100 | < ₹5 / mo3,000 bills ≈ 24 MB |
Shared across all Entyti products. Cost per pharmacy falls as shops are added — the opposite of the WhatsApp variable above.
| Item | Plan / cost | Clients it supports | Upgrade trigger → next tier |
|---|---|---|---|
| SupabasePostgres + storage + auth | Pro · $25/mo~₹2,175 org base | 100+ shops, years of runway8 GB DB ≈ 40 shop-yrs · 100 GB PDF ≈ 340 shop-yrs · now 24 MB / 10 test tenants | Not a plan change — compute add-on at ~50–100 shops billing concurrently→ Small +$15 / Medium +$60 mo |
| Cloudflare PagesSPA + /api/wa Functions | Free · $0static served unlimited | ~33 – 330 shopsFn cap 100k calls/day · 1 call ≈ 1 WA send · 300–3,000 bills/day/shop | >100k Fn calls/day→ Workers Paid $5/mo (≈333k/day) |
| AiSensyWhatsApp platform | PRO planconfirm ₹ | Unlimited pharmaciesall shops are contacts under ONE central WABA — not per-seat | SaaS plan bumps only for team seats / features, not client count |
| WhatsApp throughputMeta messaging tier | Tier 12,000 conv / 24h | ~2,000 conv/day, all shopsdashboard: 1,992 left | Auto Tier 1→2 (10k)→3 (100k) by Meta on volume + qualityfree · automatic |
| Anthropic APIAI inference | Pay-as-you-gono base fee | Unlimitedvariable only — see table A | None — scales linearly, quota-gated per shop |
| Domainsentyti.in · entyti.com | ~₹900 + ~$12per year | n/a | Renewal only |
The order you actually hit paid steps — cheapest and earliest first.
Cost to company, revenue, and profit — per shop and across the whole fleet — at each capacity milestone. Cost basis is the conservative current model (Entyti absorbs WhatsApp + AI); revenue is the cheapest tier only.
| Scale & capacity | Per shop / mo | Fleet total / mo | |||||
|---|---|---|---|---|---|---|---|
| Cost | Revenue | Profit | Margin | Cost | Revenue | Profit | |
| 10 shopsincluded tiers · ₹0 upgrades | ₹588 | ₹1,500 | ₹912 | 61% | ₹5,880 | ₹15,000 | ₹9,120 |
| 30 shopsnearing CF Functions cap (100k/day) | ₹367 | ₹1,500 | ₹1,133 | 76% | ₹11,010 | ₹45,000 | ₹33,990 |
| 100 shopsCF Paid +$5 · Supabase compute +$15 engaged | ₹307 | ₹1,500 | ₹1,193 | 80% | ₹30,700 | ₹1,50,000 | ₹1,19,300 |
| At 100-shop fleetconservative — all on Essential | 80% gross margin | ₹30,700 | ₹1,50,000 | ₹1,19,300 | |||
Revenue: Essential ₹4,499/qtr = ₹1,500/mo — the cheapest tier, taken across the board. Growth is ₹8,999/qtr (₹3,000/mo) + ₹15k one-time setup, so real profit runs higher.
Cost: avg shop = 40 WhatsApp bills/day (₹156/mo @ ₹0.13) + ~₹100 AI (quota-capped) + fixed share. Fixed = Supabase $25 (+$15 compute @100) + AiSensy PRO ~₹999 confirm + CF free→$5 + domains ~₹150. ~₹87/USD.
Suheab's usage-billing lifts margin to ~97% by moving WhatsApp+AI onto the client, dropping company cost to fixed-only (₹51/shop @100).
WhatsApp is a hard floor, not a markup. ~₹0.13/message is Meta's own India utility rate — going direct to Cloud API cuts almost nothing vs AiSensy (which already runs near cost). No official path gets below it. The only ₹0 route is free wa.me share + PDF link (one manual tap).
The two columns above are the whole decision. Eat WhatsApp yourself and company cost balloons to ~₹30,700/mo at 100 shops. Bill usage to the client (Suheab's model) and it stays ~₹5,064/mo — ₹51/shop, falling. Fixed amortizes; variable does not.
Two ₹ figures still need the billing dashboards only the owner can open: AiSensy PRO monthly fee and the actual Supabase invoice. Both sit in Fixed, so they barely move the per-shop number.
Confirmed rates in green. Soft / to-confirm figures flagged in amber. AI per-run cost is modeled, not metered — a week of token logging on the send worker would turn it hard.