What it costs to build
a food delivery app like Zomato

A one-city aggregator like an early Zomato or Swiggy costs ₹35–90 lakh; a restaurant brand's own ordering app costs ₹10–25 lakh; a multi-city marketplace costs ₹1–2.5 crore.

Customer, restaurant, rider, operations

4 apps

One-city aggregator launch

₹35–90 lakh

Design to pilot

5–8 months

A restaurant's own ordering app

₹10–25 lakh

Most founders who ask for "an app like Zomato" need one of two very different products.

If you are a restaurant group, start with your own channel. If you are building a marketplace, start in one neighbourhood with a few dozen restaurants you can personally onboard. India's ONDC network is a third path: restaurants join through a seller app and appear across several buyer apps without building a consumer app at all².

The customer app (about 25% of the build)

Search and filters, restaurant pages, menus with variants and add-ons, cart, coupons, UPI and card payments, live tracking, ratings, reorder and support chat. This is the app everyone pictures, and the one competitors have polished for a decade — so your differentiation has to come from somewhere else.

The restaurant app or dashboard (about 15–20%)

Loud order alerts that work on a kitchen tablet, accept or reject, preparation time, "mark ready", menu and stock switches ("out of paneer"), daily settlements and payouts. Restaurants judge you on this app. If an order is missed because a notification failed, they stop trusting you.

The delivery-partner app (about 20%)

Shift on/off, order assignment with timers, navigation to restaurant and customer, pickup and delivery proof (OTP or photo), cash collection where you allow it, earnings, incentives and payouts. It has to run all day on budget Android phones without draining the battery.

The operations console (about 15%)

Restaurant onboarding and document checks, menu uploads, zones and delivery fees, live order monitoring, rider management, refunds and complaints, and the numbers: orders, delays, cancellations. Without it, every problem becomes a phone call.

The remaining share is the backend that joins all four: orders, dispatch, notifications, payments and payouts.

Ride-hailing has one clock: how fast the driver arrives. Food delivery has two. The kitchen needs 15–25 minutes; the rider should arrive as the food is packed, not ten minutes early (wasted rider time) or ten minutes late (cold food, angry customer). Getting that right is the core of the product.

• Preparation-time estimates per restaurant, adjusted by real history and rush hours
• Rider assignment timed to food readiness, not to order placement
• Order batching rules: when one rider may carry two orders, and when never
• Delay detection and automatic customer updates before anyone complains
• A refund and compensation policy encoded in the system, so support is consistent
• Menu and stock changes that reach customers within seconds

Every order carries small costs that decide whether your commission or delivery fee makes sense:

• Maps. Address search, geocoding and routes for riders. Google Maps gives 10,000 free requests a month per product, then charges per call³; Ola Maps advertises a larger free tier for Indian developers⁴.
• Messages. OTPs, order updates and rider notifications. Prefer push and WhatsApp; SMS is the costly fallback.
• Payment fees on online orders, plus cash-handling costs if you accept cash on delivery.
• Rider pay per order, which is by far the largest variable cost.
• Servers that grow with concurrent orders at lunch and dinner peaks.

Write the cost of one order on a single line before you set fees. If rider pay and payment fees alone eat most of your take, no amount of app polish will fix the model.

The restaurant commission funds discovery, logistics and payments — and it is the most contested part of the business in India. Restaurant associations have long pushed back on the big platforms' terms, and in 2024 an investigation by the Competition Commission of India found that Zomato and Swiggy breached antitrust law through exclusivity arrangements with restaurant partners¹.

For a new entrant, that tension is an opening: transparent fees, faster settlements and fair menu pricing are features restaurants will switch for. Build settlement reports restaurants can understand, and make payouts fast. It costs little extra and it is the clearest way to win supply.

1. Start as a channel, not a marketplace, unless the marketplace is your edge.
2. Launch in one neighbourhood with 30–50 restaurants you onboard by hand. Density beats coverage.
3. Use a logistics partner first instead of building a rider fleet and app on day one. Bring delivery in-house once volume justifies it.
4. Build cross-platform customer apps; make the rider app Android-first.
5. Buy the commodity parts — payments, maps, OTPs, chat, push — and spend on dispatch and the restaurant experience.
6. Test with real kitchens during a lunch rush before launch. It will reveal more than any QA script.

These are Science of Apps editorial estimates, consistent with our app development cost guide, which puts a first on-demand app at ₹25–80 lakh. Food delivery aggregators sit at and above the top of that band because they add a fourth product — the restaurant app — and settlement logic for two sets of partners. Rates follow the hourly bands Indian teams publish on Clutch⁵; map pricing comes from Google's and Ola's own pages³⁴. Figures are before 18% GST and reviewed every six months.

Aggregator — a marketplace that lists many restaurants and handles discovery and delivery. Cloud kitchen — a delivery-only kitchen with no dining room. Dispatch — assigning orders to riders. Batching — one rider carrying more than one order. Settlement — paying restaurants and riders what they are owed after fees. ONDC — the Open Network for Digital Commerce, a government-backed network that connects sellers and buyer apps.