Short answer: custom app development cost in India is driven by scope, not by a price list. A small internal dashboard is a different animal from a customer portal with payments, and an Android field app that works offline is different again. What follows is how to work out your own number and how to make quotes comparable.
- Cost is driven by six things: number of screens, number of user roles, third-party integrations, offline support, payments, and app store distribution.
- Ranges quoted online are typical market ranges only. They move enormously with scope, and any firm number before a written scope is a guess.
- Running costs are separate and predictable: hosting, APIs, the Google Play one-time US$25 registration fee, Apple's US$99/year if you need iOS, payment gateway fees, and maintenance.
- The hidden costs are usually content, data migration, user training, edge cases and the second round of changes after real users touch it.
- Give every vendor the same one-page scope and you will get quotes you can actually compare.
- All third-party prices below are as of September 2026 — check the official page before budgeting.
What actually drives custom app development cost in India
When two quotes for "an app" differ by 4x, it is almost never because one firm is greedy. It is because they read the brief differently. These are the six variables that move the number most.
1. Number of screens
Screens are the crudest but most honest proxy for effort. A login, a list, a detail view and a settings page is four screens. A real product has thirty. Each screen needs design, front-end build, validation, empty states, error states and testing. When you are asked "how many screens?", the honest answer is usually more than you first thought — every form needs a success state and a failure state.
2. Number of user roles
This is the variable people underestimate the most. One role is simple. Admin plus staff plus customer is three sets of permissions, three navigation trees, three sets of test cases, and a permissions matrix that has to be right or you have a data leak. Going from one role to three often adds more work than adding ten screens.
3. Third-party integrations
Every integration is a mini-project: authentication, field mapping, rate limits, error handling, sandbox testing, and a plan for what happens when the other side is down. Tally, Zoho, a CRM, WhatsApp Business API, an SMS gateway, Google Maps, a logistics partner — each one is real work. Integrations with good public documentation are far cheaper than integrations with a PDF from a vendor's sales team.
4. Offline support
If your field staff work where there is no signal, the app must store data locally and sync later. That means conflict resolution: what happens when two people edit the same record offline? Offline capability is one of the few features that can genuinely double the cost of a mobile app, and it is worth being honest about whether you need true offline or just tolerance of a slow connection.
5. Payments
Taking money adds a gateway integration, webhook handling, reconciliation, refunds, failure states, invoices and an audit trail. It also raises the security bar for everything around it. This is not one feature; it is a subsystem.
6. App store distribution
Publishing to Google Play is not just an upload. Google charges a one-time US$25 registration fee for a developer account. Personal accounts created after 13 November 2023 must also run a closed test with a minimum of 12 testers opted in continuously for at least 14 days before they can apply for production access, and that review typically takes seven days or less. See Google's official testing requirements. Plan a few weeks for this, not a few days. For iOS, the Apple Developer Program is US$99 per membership year and is required for App Store distribution.
If your "app" is really a set of repetitive tasks that need doing automatically, you may not need an app at all — see is this task worth automating before you commission a build.
Typical ranges by build type
These are typical market ranges we see quoted in India as of September 2026, not LZ Worth prices, and not a quote. They shift with scope, seniority of the team, city, and how much design polish you want. Use them to sanity-check a quote, not to set a budget.
| Build type | Typical market range | Assumptions behind the range |
|---|---|---|
| Internal dashboard | Lower end of the market | 8–15 screens, 1–2 roles, one database, 0–2 integrations, desktop browser only, no payments, no offline, no app store, existing brand assets, staff already trained on the process. |
| Customer portal | Middle of the market | 15–30 screens, 2–3 roles, self-signup and authentication, notifications by email or WhatsApp, 2–4 integrations, payments and invoices, mobile-responsive web, basic analytics, a real design pass because customers see it. |
| Android field app | Upper end, and widest spread | 10–25 screens, 2–3 roles, offline capture with sync and conflict handling, camera and GPS, a back-office web dashboard to see the data, Play Store release, device testing across cheap Android handsets, and a rollout plan for field staff. |
Why are the rupee figures not filled in? Because a number without the assumptions is marketing, not information. The honest version is this: the same three words — "customer portal app" — can legitimately be a four-week build or a six-month build, and the price follows the weeks. What you should demand from any vendor is a range with the assumptions written next to it, exactly as in the table above. If they will not write assumptions down, the range is meaningless.
For reference on what we build in each of these categories, see custom web and Android apps and recent work.
Freelancer vs agency vs studio
| Freelancer | Agency | Small studio | |
|---|---|---|---|
| Typical cost | Lowest | Highest | In between |
| Speed to start | Fast | Slow (sales cycle, onboarding) | Fast |
| Who you talk to | The person building it | An account manager | The person building it |
| Bus factor | One. If they disappear, you are stuck. | Low risk | Moderate |
| Design quality | Varies wildly | Usually strong | Varies |
| Process and documentation | Often none | Usually formal | Depends — ask |
| Best for | Small, well-defined internal tools | Large, long-running, compliance-heavy builds | Focused builds where you want the builder accountable |
None of these is right in general. The failure modes are what matter: a freelancer who vanishes mid-build, an agency where the people who sold you the project never touch it, and a studio that overcommits. Screen for the failure mode, not the label.
Running costs after launch
Budgeting only for the build is the most common mistake. These are the recurring lines, with the verifiable third-party prices as of September 2026 — check each official page, they change.
- Hosting. A VPS for a small internal tool starts around ₹600–₹1,200 a month in India depending on specs and whether you are on a promotional or renewal rate. Managed cloud costs more and needs less of your attention.
- Domain and SSL. A domain is a few hundred to about two thousand rupees a year; SSL is free with Let's Encrypt on most stacks.
- Google Play. A one-time US$25 registration fee for the developer account. Not recurring.
- Apple App Store. US$99 per membership year, required for App Store distribution. Recurring, and it does recur — miss it and your app comes down.
- Payment gateway. Razorpay's published standard rate is about 2% plus 18% GST on the fee for most domestic methods including cards, UPI, netbanking and wallets, and about 3% plus GST for international cards, with no setup fee or annual maintenance charge on the standard plan. Confirm on Razorpay's official pricing page, because rates and method-wise exceptions change.
- Transactional messaging. SMS, email and WhatsApp Business API conversations are per-message or per-conversation. At low volume this is small; at scale it can exceed hosting.
- Third-party APIs. Maps, OCR, LLM calls, address lookup, credit checks. Meter them from day one or you will be surprised.
- Maintenance. Budget a percentage of the build cost annually for OS updates, dependency upgrades, security patches and small changes. Android and browser versions move whether you touch the app or not.
Hidden costs nobody quotes for
- Content and data entry. Somebody has to write the labels, help text, terms, privacy policy and email templates. That somebody is usually you.
- Data migration. Getting ten years of messy spreadsheets into a clean schema is often the single biggest unplanned task in a build.
- Training and rollout. A field app that staff will not use is a total loss. Budget for training sessions, a one-page guide and a support person for the first month.
- Change requests after first contact with reality. Real users always want changes. Reserve roughly 15–20% of the build budget for round two rather than pretending the first version is final.
- Edge cases. Refunds, cancellations, partial payments, duplicate records, staff who leave mid-process. These are where the last 20% of the effort hides.
- Testing on cheap Android devices. Your app has to work on a ₹9,000 phone on a 3G connection, not just on the developer's handset.
- Compliance and legal review. If you handle customer personal data, someone should read your consent flows and retention policy.
- Handover. Source code, credentials, deployment access, documentation. If this is not in the contract, it is not free.
How to scope so quotes are comparable
The single highest-return hour you can spend is writing one page that every vendor sees. Include exactly this:
- The problem in one paragraph. What happens today, manually, and what it costs you in hours or errors.
- The user roles. Name them. For each, one line on what they can see and do.
- The screen list. Just names. "Login, dashboard, job list, job detail, photo upload, sync status, admin user list, reports."
- Integrations. Name each system, and whether you already have API access and documentation.
- Must-haves vs nice-to-haves. Two columns. Be ruthless. This is where cost gets controlled.
- Volumes. Users, records per month, peak concurrency. Ten users and ten thousand users are different builds.
- Platform. Web only, Android only, both, iOS needed or not.
- Offline. Genuinely required, or just nice? Say which.
- Deadline and why. A real constraint, not "ASAP".
- What you will provide. Brand assets, content, test data, a named decision-maker.
Then ask every vendor to quote against that page, in phases, with a fixed price and a written scope per phase. Ask each of them to list what they have excluded. The exclusions tell you more than the inclusions.
Red flags in a quote
- A single number with no scope attached. If the quote is one line, the argument is coming later.
- No exclusions listed. Every honest quote excludes something.
- "Unlimited revisions." Nobody means this. It signals either a vague scope or a plan to bill you elsewhere.
- Hourly-only pricing with no cap on a project where you cannot judge the hours.
- The vendor owning the code, repository or accounts. You should own the source code, the server, the domain and every API key. If a vendor resists this, walk.
- No mention of testing, handover or documentation anywhere in the proposal.
- A quote that dropped by 40% when you pushed back. That means the first one was arbitrary, and so is the second.
- Payment schedule loaded to the front. Milestones should be tied to delivered, working things.
- No named developer. Ask who is actually writing the code and whether you can talk to them.
- Timelines with no dependencies on you. Real plans say "we need your content by week two".
A sensible way to start
Do not commission the whole thing. Commission the smallest version that a real user can use for a real job, get it into their hands, then decide what round two is with evidence instead of opinions. Almost every expensive app failure we have seen started as a fully specified six-month build that nobody used.
Also check whether an app is the right tool at all. A lot of what businesses ask apps for — chasing approvals, generating reports, sending reminders, copying data between systems — is automation work, not product work, and costs a fraction as much. Our writeups on automation ideas for business and AI automation cover that side.
If you want a fixed scope and price in writing, that is how we work: a free 30-minute call, then scope and price agreed in writing within 48 hours, and you own all code, keys and accounts at handover. Start at services or read more on the blog.
Frequently asked questions
How much does it cost to build a custom app in India?
There is no single answer, and any vendor giving one before seeing a scope is guessing. Cost tracks the number of screens, user roles, integrations, whether it works offline, whether it takes payments, and whether it ships to an app store. An internal dashboard sits at the low end of the market, a customer portal in the middle, and an offline Android field app at the top. Ask for a range with the assumptions written next to it.
Why do two quotes for the same app differ so much?
Almost always because the two vendors assumed different scopes. One counted eight screens, the other twenty-two. One assumed you already have API documentation for your CRM, the other budgeted to reverse-engineer it. Give both the same one-page scope, ask both to list exclusions, and the gap usually collapses.
What does it cost to publish on the Play Store?
Google charges a one-time US$25 registration fee for a developer account. Personal accounts created after 13 November 2023 must also run a closed test with at least 12 testers opted in continuously for 14 days before applying for production access, and that review usually takes seven days or less. Budget the calendar time, not just the fee. Prices as of September 2026 — check Google's official page.
What are the ongoing costs after launch?
Hosting from roughly ₹600–₹1,200 a month for a small VPS in India, a domain, payment gateway fees (Razorpay's published standard rate is about 2% plus GST for most domestic methods), transactional SMS, email or WhatsApp charges, any third-party API usage, Apple's US$99 per year if you ship to iOS, and an annual maintenance allowance. All as of September 2026.
Should I hire a freelancer or an agency?
It depends on the failure mode you can tolerate. A freelancer is cheapest and fastest to start but carries single-person risk and often no documentation. An agency has process and continuity but costs more and puts an account manager between you and the builder. A small studio sits in between. For a small, well-defined internal tool, a good freelancer is often the right call.
How much should I keep aside for changes after launch?
Reserve roughly 15–20% of the build budget for the first round of changes after real users touch the app. This is not a contingency for vendor mistakes; it is the normal cost of learning what users actually need, and planning for it stops it becoming an argument.
Who should own the code and the accounts?
You should — the source code repository, the server, the domain, the app store listing and every API key. A vendor holding any of these is a commercial lock-in, not a technical necessity. Put ownership and handover in the contract before work starts, and ask for a recorded walkthrough at handover so your team can operate it.
Deepanshu Singh co-founded LZ Worth, a Delhi studio that builds AI automations, apps and content systems. He has tested more than 150 AI tools and would rather tell you which ones to skip. More about him.