Running this platform costs under ₹10,000 a month at launch and under ₹75,000 a month at thirty times the volume. Year-one cash comes in around ₹3 lakh.
The expensive resource is a developer's year — roughly seven to fourteen times the entire cash cost. Every number below is built up from vendor list prices so you can challenge any line.
Three scenarios: launch in one locality, month twelve across three or four localities, and a mature single city. Platform cost barely moves; per-booking costs are what grow.
Masked calling goes from 19% of the bill at launch to 33% at scale, overtaking everything else as the single largest line. It's also the one line that grows strictly with booking volume and cannot be optimised by better engineering. This is the concrete reason in-app chat is on the v2 list — shifting even half of call minutes to chat is worth roughly ₹12,000/month at 10,000 bookings, which is more than the entire platform bill at launch.
Vendor list prices as of August 2026, converted at ₹95.3 / USD (the August 2026 monthly average). GST at 18% is included where it applies to Indian vendors.
| Item | Basis | 300 bookings | 3,000 | 10,000 |
|---|---|---|---|---|
| Platform | ||||
| Supabase | Pro $25 + compute add-on at scalePostgres, auth, storage, realtime, PostGIS | ₹2,375 | ₹4,375 | ₹9,500 |
| Vercel | Pro $20 + function usageweb, API, admin — all one deploy | ₹1,900 | ₹3,400 | ₹6,000 |
| Expo EAS | Free tier → Production $29builds + over-the-air updates | ₹0 | ₹2,755 | ₹2,755 |
| Cloudflare R2 + Images | Storage + egressjob photos, parts bills | ₹250 | ₹800 | ₹2,500 |
| Upstash QStash | Free tier → paiddispatch timeouts | ₹0 | ₹500 | ₹1,500 |
| Sentry + PostHog | Free tiers → paiderrors, analytics, session replay | ₹0 | ₹3,900 | ₹8,500 |
| Domain | Amortised annual | ₹125 | ₹125 | ₹125 |
| Platform subtotal | ₹4,650 | ₹15,855 | ₹30,880 | |
| Messaging | ||||
| WhatsApp messages | ≈8.3 utility msgs/booking @ ₹0.115OTP, 3 job offers, accepted, en route, estimate, agreed, complete | ₹290 | ₹2,875 | ₹9,545 |
| BSP platform fee | AiSensy / Interakt tierIndia-registered WABA — mandatory | ₹1,800 | ₹2,500 | ₹4,000 |
| Push + SMS fallback | Expo Push free; SMS last resort | ₹310 | ₹925 | ₹2,455 |
| Messaging subtotal | ₹2,400 | ₹6,300 | ₹16,000 | |
| Voice | ||||
| Masked call minutes | ≈1.5 calls × 3 min/booking @ ₹0.50Exotel bridge — privacy and anti-leakage | ₹675 | ₹6,750 | ₹22,500 |
| Virtual number rental | Pool scales with concurrency | ₹1,025 | ₹1,500 | ₹2,000 |
| Voice subtotal | ₹1,700 | ₹8,250 | ₹24,500 | |
| Other | ||||
| Maps | Mappls free tierno map SDK needed in v1 | ₹0 | ₹400 | ₹1,200 |
| Misc / buffer | Email, backups, small tools | ₹250 | ₹800 | ₹1,400 |
| Other subtotal | ₹250 | ₹1,200 | ₹2,600 | |
| Total per month | ₹9,000 | ₹31,600 | ₹74,000 | |
| Per booking | ₹30.00 | ₹10.53 | ₹7.40 | |
Start the company registration early — it takes two to four weeks and gates the WhatsApp Business API, the payment gateway and Play Store publishing.
| Item | Low | High |
|---|---|---|
| Private Limited registrationgovernment fees + professional | ₹10,000 | ₹20,000 |
| GST registrationfiling is free; professional fee only | ₹1,500 | ₹3,000 |
| Apple Developer Program$99 / year | ₹9,400 | ₹9,400 |
| Google Play Console$25 one-time | ₹2,400 | ₹2,400 |
| Domain, first year | ₹1,000 | ₹1,500 |
| WhatsApp BSP onboarding | ₹0 | ₹5,000 |
| Trademark filingoptional, but cheap insurance on the name | ₹9,000 | ₹15,000 |
| Total | ₹33,300 | ₹56,300 |
Background checks run ₹150–400 per provider through AuthBridge or IDfy. Fifty providers at launch is ₹7,500–20,000; 150 across year one is ₹22,500–60,000. It scales with supply, not bookings — so it front-loads in each new locality.
Setup ₹45,000 · infra months 1–6 at ₹9,000 ≈ ₹54,000 · infra months 7–12 averaging ₹20,000 ≈ ₹1,20,000 · 150 provider checks ≈ ₹40,000 · 15% contingency ≈ ₹39,000.
≈ ₹3.0 lakh, with a sensible range of ₹2.5–3.5 lakh.
A cost model that stops at cloud bills is misleading. For a solo build, developer time is the dominant input by an order of magnitude, and it should be on the same page as the ₹9,000.
If time is 7–14× the cash, then every decision should buy time with money, never the reverse. That is the entire justification for managed services: ₹2,375/month for Supabase replaces roughly 40% of the backend work, which at these rates pays for itself in the first three days of the year. Self-hosting on a ₹1,500/month VPS would save ₹10,000 a year and cost weeks — a straightforwardly bad trade at n=1.
It is also the argument for the v1 cut list. Every feature deferred is time returned to the only genuinely scarce resource.
In rough order of impact, so you know which assumptions to challenge first.
The model assumes 1.5 calls × 3 minutes. If real usage is 3 calls × 4 minutes, the voice line roughly triples and the 10,000-booking total goes past ₹1.2 lakh/month. In-app chat in v2 is the mitigation, and this is why it ranks above almost everything else on that list.
Assumed 8.3 utility messages. Batching status updates, or leaning on the free 24-hour service window after a customer replies, could cut this by a third. Conversely, careless marketing messages at ₹0.86 — roughly 7½× the utility rate — could dominate the entire bill.
Around 40% of the monthly bill is USD-denominated. The rupee moved from ₹90.7 in January 2026 to ₹95.3 in August — a 5% shift, worth about 2% on the total. Worth re-checking, not worth hedging at this size.
The step from the flat Pro plan to a compute add-on is the one genuine platform cliff in the model. It arrives when concurrent connections and PostGIS queries outgrow the shared instance — realistically somewhere between 2,000 and 4,000 bookings a month.
₹150–400 each, front-loaded when entering a new locality. Aggressive multi-city expansion turns this from a rounding error into a real line — 500 providers is ₹75,000–2,00,000.
Not in this model, because online payments are deferred. When they arrive, expect roughly 2% + GST on transaction value — a cost that scales with GMV and is normally netted against the take rate rather than paid from this budget.
Illustrative arithmetic against the monetisation ladder, not a forecast — the take-up rates below are assumptions, not evidence.
Revenue ₹0. Cost ≈ ₹9,000/month. You are buying liquidity with your own time, which is the correct trade at this stage.
120 active providers × 30% take-up × ₹499 ≈ ₹18,000/month. Roughly covers infrastructure at that volume. Zero friction for customers.
25% of 3,000 bookings × ₹450 × 13% ≈ ₹44,000/month. Cash stays at 0% commission permanently.
Cash breakeven is not the interesting question — at these amounts it arrives early and easily. The question that decides whether this business exists is the 60-day customer repeat rate. Homejoy reached 31 cities and still failed, because customers found a good worker and hired them off-platform. If repeat sits below 25%, no cost model on this page matters.
These are modelling assumptions for planning, not financial advice or a forecast. Verify every vendor price directly before committing to contracts.