NM Company
A portfolio and enquiry site for events firm NM Company
A visual portfolio and enquiry website for an event management and supplies firm — presenting a full...
Read itWeb application engineering
Web Development
Production web applications built on Laravel, Next.js and Postgres — typed, tested, observable, and fast on the connections your customers actually have.
- Sub-2s Largest Contentful Paint on a mid-range Android over 4G
- Typed API contracts — the front end cannot silently drift from the backend
- Every deploy reversible in under 60 seconds
- You own the repository, the pipeline and the infrastructure on day one
What this is
The website is the product. Build it like one.
Most of the web work that lands on our desk is a rescue. A marketing site that took eleven seconds to render on a 4G connection in Indore. A booking flow where two customers could pay for the same slot. A CMS so tangled the client's own content team had stopped using it and gone back to emailing PDFs.
We build web applications the way software gets built, not the way brochures get assembled. Typed contracts between the front end and the API. Migrations under version control. A test suite that runs on every pull request. Deploys that take four minutes and can be rolled back in one. The visible half is the interface; the half that decides whether you are still happy in year three is everything underneath it.
This is the right service for you if the site is the product — a marketplace, a portal, a dashboard, a booking engine, a SaaS front end, a content platform with real editorial workflow. If you want a five-page company profile, tell us and we will say so honestly rather than sell you a project you do not need.
Sound familiar
The problems people bring us.
Not abstractions. These are the exact sentences that arrive in our inbox.
-
The site takes 9 seconds to load on a phone
Three megabytes of unused JavaScript, hero images shipped at 4000px wide, and a homepage that fires 140 database queries. Your bounce rate is not a marketing problem.
-
Nobody knows what the data actually means
Two tables both called users. Order status stored as a free-text string with eleven distinct spellings. Reporting is guesswork because the schema was never designed, only accumulated.
-
Deploys happen at midnight and everyone holds their breath
Files dragged over FTP. No staging environment. No rollback. The last person who understood the server left in 2023 and took the SSH key with them.
-
The security audit came back ugly
Admin endpoints protected by nothing but an unguessable URL. Passwords hashed with MD5. File uploads written straight into the web root. A DPDP Act notice you cannot honestly answer.
-
The content team gave up on the CMS
Adding a case study requires a developer. The page builder produces layouts that break on mobile. Your marketing calendar is now bottlenecked on somebody else's sprint.
-
Nothing talks to anything else
The website, the CRM and the accounting system each hold a different version of the customer. Somebody reconciles them by hand in a spreadsheet every Friday.
What good looks like
Numbers we hold ourselves to.
Every build carries a performance and reliability budget from day one, enforced in continuous integration. These are the medians across the web applications we have shipped in the last three years — not a best case we found in an old invoice.
See how we work
0s
Median LCP
Throttled 4G, mid-range Android
0+
Lighthouse performance
Measured on production, not localhost
0min
Deploy time
Commit to live, fully automated
0%
Uptime
Rolling 12-month average
How we build
Five phases. No surprises in month three.
Every phase ends in something you can look at, click on, or read — never a status update.
-
01
Discovery and technical shaping
1–2 weeksWe interview the people who will use the thing, not just the people paying for it. We map the domain, draw the data model, list the integrations and name the risks out loud. You leave with a written architecture note, a scope with edges, and a cost we will stand behind. If the honest answer is that your budget does not fit the ambition, you hear it here — before anyone has written a line of code.
-
02
Interface and design system
2–3 weeksWireframes for the flows that carry money or risk, then a component library in Figma — tokens for colour, type and spacing, states for every interactive element, and the empty, loading and error cases nobody remembers to draw. We prototype the two or three screens where the product actually lives and put them in front of real users before we commit them to code.
-
03
Engineering in two-week increments
6–14 weeksVertical slices, not layers: each sprint delivers a working path from interface to database, deployed to a staging URL you can hit from your phone. Every change goes through a pull request with a named reviewer. The test suite, static analysis and the performance budget all run in CI. You get a demo and a written changelog every second Friday.
-
04
Hardening and launch
1–2 weeksLoad testing against realistic traffic, a security pass covering the OWASP top ten, accessibility remediation to WCAG 2.2 AA, and a dress rehearsal of the cutover on a clone of production. We agree a rollback plan before we agree a launch date. Go-live happens on a Tuesday morning, with the whole team watching the dashboards — never on a Friday evening.
-
05
Hypercare and handover
30 days, then ongoingFor the first month after launch we treat your error rate as our error rate. Defects are fixed at no charge. We tune the queries and cache layers that only real traffic reveals. Then we hand over runbooks, an architecture decision log and an onboarding session for your team — or roll into a retainer if you would rather we kept the pager.
What you get
Everything in the box.
Not upsells. This is what a web application needs to survive contact with real users, so it is in every build we do.
A performance budget in CI
Bundle size, query count and LCP thresholds enforced on every pull request. A regression fails the build; it does not quietly ship.
A typed API contract
OpenAPI or tRPC between front end and backend, with types generated from the source. Rename a field and the compiler tells you what broke.
Auth and authorisation done properly
Role and policy-based access enforced server-side, session hardening, rate limiting, and an audit trail on every privileged action.
A CMS your editors will use
Structured content, live preview, draft and scheduled publishing, media library with automatic image derivatives. No page-builder spaghetti.
Technical SEO that holds up
Server-rendered HTML, clean canonical and hreflang handling, JSON-LD schema, an XML sitemap that reflects reality, and Core Web Vitals in the green.
Observability from day one
Structured logs, error tracking with source maps, uptime checks and a latency dashboard. When something breaks you will know before your customers tell you.
The stack
Boring, proven, hireable.
We choose tools with a decade of momentum behind them and a hiring pool in India — so you can staff this internally later without a rescue mission.
Laravel 12
PHP 8.3
Node.js
Livewire
Inertia.js
Laravel Horizon
Next.js
React
TypeScript
Tailwind CSS
Alpine.js
Vite
PostgreSQL
MySQL
Redis
Meilisearch
S3 / Cloudflare R2
GitHub Actions
Docker
Terraform
Playwright
PHPStan
Sentry
Performance and security are not a phase.
Performance is not a garnish you add at the end. It is an architectural decision you make on day one and then defend on every pull request. We set a performance budget before a single component is written — a JavaScript budget in kilobytes, a Largest Contentful Paint target in milliseconds, a limit on the number of database queries any single request may fire — and CI fails the build when a change blows through it.
The same goes for accessibility and security. Keyboard traps, unlabelled form fields, colour contrast below 4.5:1, an admin route without an authorisation policy: these are bugs, and they get caught in review, not in an audit six months after launch when a legal notice arrives.
The difference
Why this costs more than a template shop.
| Template shop | PRS India | |
|---|---|---|
| Who writes your code | A junior on four projects at once | A senior engineer with a named reviewer |
| Automated tests | No | Yes |
| Staging environment | No | Yes |
| Rollback plan | Restore last week's backup | One command, under 60 seconds |
| Performance budget enforced in CI | No | Yes |
| You own the repo and cloud account | Sometimes, after a fight | From the first commit |
| What happens in year three | Rewrite | Still shipping features |
Proof
All case studies
Work that is still in production.
Real systems, real load, real numbers — with the constraints and trade-offs written down.
FealDeal
E-commerce — Ethnic Fashion
A high-catalogue ethnic-wear storefront for FealDeal
A Shopify storefront for a Surat ethnic-wear label — 200+ SKUs of lehengas, sarees and suits organised by type, colour, fabric and work, wit...
Shopify
Platform
0+ SKUs
Catalogue
Turant Logistics
Logistics & Supply Chain
Cutting order-to-dispatch by 62% for a 400-truck 3PL
A fleet, ePOD and route-optimisation platform built offline-first for drivers in low-connectivity corridors — because the two off-the-shelf...
0% faster
Faster order-to-dispatch on the ops floor
0% shorter
Shorter POD-to-invoice cycle (8.4 days to 19 hours)
Collection by Rehan
E-commerce — Watches & Accessories
A brand-led watch and eyewear store for Collection by Rehan
A Shopify storefront for a Surat watch and eyewear retailer — browsable by gender, brand, style and combo, with predictive search, COD check...
Shopify
Platform
Brand & style
Browse by
A focused product — one core workflow, an admin, auth and payments — is typically 10 to 14 weeks from kickoff to production. A multi-role platform with integrations into an ERP, a payment gateway and a logistics partner runs 16 to 24 weeks. We scope in two-week increments and ship something demonstrable at the end of each one, so you are never waiting three months to find out whether the thing works.
By what the application actually does. Heavy server-side logic, complex permissions, back-office tooling, payments and reporting: Laravel, usually with Livewire or Inertia so we are not maintaining two codebases for no reason. Content-heavy, SEO-critical, or a rich interactive front end over an existing API: Next.js. Many of our builds are both — Laravel serving a typed API, Next.js rendering the public surface. We make the call in week one and write down why.
That is the only performance target that matters in India. We test on throttled 4G against a mid-range Android device, not on a MacBook over office fibre. That means server-rendered HTML where it counts, images served as AVIF or WebP at the right dimensions, a JavaScript budget enforced in CI, and a CDN edge in Mumbai and Chennai. If a page cannot hit our LCP target on that profile, it does not ship.
Completely, from the first commit. The repository lives in your GitHub or GitLab organisation, the cloud account is in your name and billed to you, and every credential is handed over in a password manager vault you control. There is no proprietary framework, no licence fee and nothing that stops another firm from picking the project up tomorrow. We would rather earn the retainer than trap you into it.
A 30-day hypercare window is included in every fixed-scope build: we watch the error rate and the latency graphs, fix defects at no charge, and tune the parts that only real traffic exposes. After that you can take it in-house — we hand over runbooks and an architecture document — or keep us on a support retainer with a defined response SLA and a monthly pool of engineering hours.
Regularly. We start with a paid two-week audit: we read the code, run it locally, map the data model, and hand you a written report on architecture, security exposure, dependency rot and test coverage, with a prioritised remediation plan and honest costs. Sometimes the conclusion is that the codebase is worth investing in. Sometimes it is that a rewrite is cheaper than the maintenance. You get the truth either way.
Send us the version of the problem you think is unbuildable.
A senior engineer reads every enquiry and replies within one business day with scope, risk and a number. No discovery call required to get a straight answer.
Proof
Shipped, measured, still running.
Swasthya Sarathi
A healthcare-companion platform for Swasthya Sarathi
A multi-service healthcare platform that helps people find hospitals, doctors, labs, medical stores,...
Read itMV Tech Education
A course and admissions platform for MV Tech Education
A course-catalogue and admissions website for a Bihar vocational institute — leading with industrial...
Read itRelated services
Mobile App Development
Flutter, React Native and native apps that hold up offline, on old hardware and on a metro...
ExploreEcommerce Development
Commerce platforms where the checkout converts and the back office reconciles — multi-ware...
ExploreERP Development
ERP systems built on an append-only stock and financial ledger — multi-plant, multi-GSTIN,...
ExploreLet's talk web development
Bring us the hard version of the problem.
A senior engineer reads every enquiry. You'll get a real answer — scope, risk and a number — within one business day.