Every question we've actually been asked.

For most marketing sites, WordPress is overkill. It's a CMS designed for 2008's publishing workflow. Modern static sites are faster, more secure, cheaper to host, and easier to maintain — if you know how to build them.

We wire up a headless CMS (Sanity or Contentful) so your marketing team can edit copy, swap images, and publish changes without touching code. No plugins. No PHP errors. No "fatal error" at 2AM.

Static sites win SEO by default — fast, semantic HTML, clean URLs, proper sitemaps. We configure structured data, meta tags, and Core Web Vitals so you're competing on content quality, not technical debt.

Yes. We've migrated from WordPress, Wix, Squarespace, Webflow, and hand-coded PHP. We preserve URLs and set up redirects so you don't lose Google rankings.

A marketing site with 8-15 pages typically takes 4-5 weeks from kick-off to launch. Landing pages: 2 weeks. Complex sites with custom functionality: 6-8 weeks.

We'll deploy to Vercel, Netlify, or Cloudflare — whichever makes sense for your traffic. Hosting costs are typically ₹0–₹2,000/month for marketing sites. You own the account.

Depends on your volume and complexity. Under 500 orders/day with standard products: Shopify. Complex inventory or B2B: custom. We give you an honest recommendation, not whichever we earn more commission on.

We implement OTP verification before order confirmation, pincode-level COD toggles, and automated risk scoring. Our clients see 60-70% reduction in fake COD orders.

Every order generates an invoice with your GSTIN, HSN codes, tax breakup, and e-invoice number (if applicable). Auto-emailed to the customer, auto-synced to your accounting software.

Yes. We handle theme migration, app re-implementation, customer data transfer, and URL preservation for SEO. Minimum downtime — typically zero.

Headless is worth it if you're doing over 5,000 orders/day, need custom checkout flows, or want to unify web + app + marketplace. For smaller stores, it's premature optimisation.

Shiprocket, Delhivery, Bluedart, DTDC, India Post, Shadowfax — we've integrated all of them. Automated AWB generation, tracking pages, return workflows.

Yes. Most engagements involve your internal engineers. We pair-program, review PRs, run architecture workshops, and eventually hand off. We're not trying to own your codebase forever.

Standard NDA signed before first architecture call. All IP transfers to you on delivery. Code is in your GitHub org from day one. We never retain any proprietary code.

Monitored, logged, load-tested, documented. CI/CD pipeline. Rollback strategy. Incident response playbook. Not "it works on staging, ship it."

Yes. We've migrated Django monoliths to microservices, PHP to Node.js, .NET Framework to .NET Core. Strangler fig pattern — no big-bang rewrites.

For new web applications, we only take projects with at least a 10-week scope. Anything smaller usually means the problem isn't scoped yet.

For v1, yes — fixed scope, fixed timeline, milestone-based. Post-launch iteration, we move to a retainer. Mixing these is where most projects break.

For a 5-person sales team doing email-first B2B, Zoho is fine. If you're WhatsApp-first, have unusual lifecycle stages, or want deep call-tracking integration with Indian telephony (Exotel, Ozonetel, Knowlarity) — custom is often cheaper long-term and definitely faster for your team.

Yes. We pull leads from IndiaMART, Justdial, Sulekha, Magicbricks, 99acres, Housing, and any Meta / Google Ads form. Deduplication, auto-assignment, first-response SLA — all handled.

Official only. Unofficial WhatsApp automation gets your business number banned. We set you up on Cloud API via Meta's official BSP partners.

Always. Built-in CSV export, SQL access on request, scheduled email exports. Your data is yours — no vendor lock-in.

Typically ₹20K–₹60K/month depending on usage. Includes hosting, support, minor features. Major new features are quoted separately.

For sales teams, we always ship a native mobile app (React Native). Your field team isn't going to open Chrome to log a call.

For under 500 crore turnover, SAP is overkill — licence fees alone run higher than our entire build. NetSuite works but forces US-centric workflows on Indian operations. For 50–500 Cr businesses, custom often wins on 3-year TCO.

Tally is an accounting tool pretending to be an ERP. It works until you cross about ₹30 Cr turnover or 3 locations — then the plugins start fighting each other.

We do. Every GST rule change since 2017 has been shipped to our clients as a non-billable update. It's part of the retainer.

Phase 1 (sales + inventory + GST): 16-20 weeks. Full ERP rollout: 9-14 months. Anyone promising less is selling you Tally with a new skin.

Migrated, validated, and kept accessible for read-only reference for 2 years. Full ledger history preserved.

You do. Code lives in your GitHub org from day one. If you want to engage another vendor in year 3, you can.

A single landing page: 10 business days from brief to live. Faster if copy is already decided.

No. We build the page; your media team / agency runs traffic. We work well with both.

Yes. We wire up a simple CMS so your team can change headlines, update proof, swap images without a developer.

A/B test winner promotion every 2 weeks, new proof updates monthly, seasonal refreshes quarterly. Usually 2-4 hours/month per page.

Yes, conversion-focused copy is included. Brand tone / storytelling content is separate.

Every major CRM, email platform, and marketing automation tool. We map form fields to your lead stages.

SharePoint works if your needs are 80% document storage. If you need HR workflows, dashboards, self-service, or partner/vendor access — custom is almost always better for the same 3-year cost.

Yes. SSO via Azure AD, Okta, Keycloak, or Google Workspace. Users don't get another password to forget.

All portals are built on audited infrastructure. Role-based access, audit logs, data encryption at rest and transit. We've supported clients through both certifications.

We wire up a friendly CMS where HR, marketing, and department leads can publish without IT. Version history, approval workflows, scheduled publishing.

Typically ₹30K-₹1.2L/month depending on user count and modules. Includes hosting, support, minor features, security updates.

If your app needs Bluetooth, advanced camera, background location, or heavy offline DB — native. If it's content, e-commerce, booking, or internal tools — PWA. We'll tell you honestly after a 30-min call.

Yes, as a TWA (Trusted Web Activity). Play Store distribution with a PWA codebase. iOS App Store is harder — Apple still restricts this.

Fully on Android and desktop. iOS added support in iOS 16.4+ — works, but with more restrictions than native.

Yes. That's the whole point. Browser prompts "Add to Home Screen" after a few visits, or you trigger it programmatically.

PWAs are still web apps — Google indexes them fully. Better than most native apps, which are invisible to search.

Roughly 45-60% of the cost of building separate iOS + Android apps. And ongoing maintenance is one codebase, not three.

For public APIs consumed by many external parties: REST + OpenAPI. For internal APIs serving multiple first-party clients (web, mobile, admin): GraphQL or tRPC. Both have their place.

URL versioning for public APIs (/v1, /v2). Header versioning for internal APIs where we control clients. Deprecation policy is always explicit — minimum 6 months notice.

We've written middleware layers for all of these. Typically an event-driven integration that keeps systems in sync without brittle point-to-point coupling.

Always. Auto-generated from schema, hosted on your domain, with authentication, changelog, SDKs, and code samples. Not a Word doc attached to an email.

Unit tests for business logic, contract tests for schema compliance, integration tests with real DBs, load tests before launch. Minimum 80% coverage on critical paths.

We never break existing clients silently. Breaking changes go through versioning, deprecation warnings, and a migration window. Clients fail loudly or not at all — never confusingly.

Swift, always, for iOS-first products or any app where UX quality is a differentiator. React Native is fine when iOS is the secondary platform and budget is the primary constraint.

Yes — account setup, provisioning, screenshots, preview videos, metadata, in-app purchase configuration. Everything up to and including responding to review feedback.

Our first-submission approval rate is ~95%. If it does get rejected, we handle the resubmission as part of the project — no extra cost.

Yes. Universal apps where it makes sense. Or iPhone-only if an iPad version would dilute the experience.

VoiceOver, Dynamic Type, and WCAG 2.1 AA compliance are default. Not optional, not afterthought.

We run a 90-day warranty. Crash-free session targets. On-call for critical issues. Optional monthly retainer after.

Kotlin, always, for new builds. Java codebases we happily maintain, but new development is Kotlin — it's more concise, safer, and officially recommended by Google.

Compose for new apps. XML for legacy we maintain. Compose is significantly faster to develop in and easier to hire for going forward.

Yes, end-to-end. Console setup, policy compliance, listing, screenshots, rollout strategy, and responding to any policy flags.

All our apps are shipped as App Bundles. We use Play Asset Delivery for apps where initial APK size matters (games, content-heavy).

Yes. Enterprise apps are built MDM-ready. App config, per-app VPN, managed configurations — all supported.

We typically target Android 7.0+ (95% device coverage). Going older (5.0, 6.0) is possible but rarely worth it — those users rarely transact.

Expo for 95% of apps — faster development, easier CI, managed native builds via EAS. Bare RN when you need custom native modules Expo doesn't support.

For UI-heavy apps with Reanimated 3: identical. For CPU-heavy apps (video editing, games, AR): native wins. For most business apps: RN is indistinguishable.

Yes. We write Swift/Kotlin modules when the JS bridge is too slow or a feature isn't available in RN. This is how the best RN apps are actually built.

Yes, via EAS Update or CodePush. Fix bugs without App Store review. Policy-compliant — we only push JS changes, not feature changes that require store review.

All new projects ship on the New Architecture by default. Fabric for rendering, TurboModules for native. It's stable as of late 2024.

Largely yes. Hooks, JSX, state — identical. They'll need to learn navigation, styling, and platform specifics, but the core mental model carries over.

Flutter if UI polish is critical, animations are heavy, or you want desktop / web from the same codebase. RN if your team is already React, or your app is CRUD + navigation heavy.

Less than it used to be. There's a healthy Flutter community in India. If you're in Bangalore, Pune, Hyderabad — hiring Flutter devs is fine. Delhi / Mumbai, slightly harder.

Ready for internal tools and dashboards. Not ready for public consumer web — SEO and initial load are still weak. We use Flutter web for B2B; Next.js for B2C.

Google has reaffirmed commitment repeatedly. Used internally in Google Pay, Google Classroom, Stadia. Community is strong. It's not going anywhere.

Yes, via platform channels or Pigeon. We do this for a couple of features in most apps — usually deep OS integrations.

Flutter adds ~4-6MB to your APK for the engine. Negligible on modern devices; noticeable on ₹6K entry-level phones. We optimise where it matters.

RN or Flutter is fine for commerce. Native is overkill unless you have >5M users or very custom interactions (AR try-on, complex video).

Yes, typically. Mobile web catches first-time visitors from Google and ads. App captures repeat users. They serve different lifecycle stages.

Branch.io or Firebase Dynamic Links. Your ad clicks, SMS links, and WhatsApp shares all land users at the exact product, not the home page.

Algolia or Meilisearch. Instant search, typo tolerance, personalised results. The default "react-native search" is bad UX.

Yes. AppsFlyer, Adjust, or Singular — we've integrated all three. SKAdNetwork on iOS, Play Install Referrer on Android.

We target API 24+ (Android 7.0) — 96% coverage including old phones. Test on Redmi 6A, Moto E, budget Samsung. Use WatermelonDB for fast local queries even on 2GB RAM.

Yes. We integrate with MDMs for zero-touch enrolment. The rep gets the phone, scans a QR, and the app + configs are pre-installed.

Via MDM (Intune, MobileIron), yes. Managed kiosk mode, app whitelisting, DLP policies — all supported.

Yes. Check-in/out with geo + time stamps. Exception reporting for out-of-bounds activity. Battery-optimised location tracking.

Daily / real-time sync via middleware. We've integrated SAP ECC, S/4HANA, Oracle EBS, Tally Prime. Order → Sales Order → Invoice automated.

Photos auto-compressed to 800KB. Voice notes optional. All synced in background when bandwidth allows.

Depends. GPT-4o for conversational breadth. Claude for nuanced multi-turn. Gemini for cost at scale. We benchmark all three on your actual data before picking.

RAG grounded in your documents, strict output constraints, confidence thresholds, and human-in-the-loop for low-confidence answers.

Sensitive data stays on our infrastructure. We use LLM APIs with no-training clauses and on-prem options (Llama 3, Mistral) for highly regulated use cases.

Yes. We train on Indian-language data and post-process for natural expression. Not word-by-word Google Translate output — actual fluent conversation.

Most projects run 10-18 weeks from kickoff to production launch. We share a milestone plan in week one and update weekly.

Yes. Standard mutual NDA on request, before the first technical conversation.

You do. Code is in your GitHub org from day one. All IP transfers unambiguously on delivery.

For v1 builds: fixed scope, fixed milestones. For ongoing work: monthly retainer with a defined team. We don't do time-and-material surprise billing.

Translate is a translator. It doesn't understand. For intent classification, entity extraction, sentiment — you need models trained on the language, not post-translation English.

Yes. Tanglish ("vanga saapdalam") and Hinglish ("kal milte hain") are first-class. We train on mixed-script corpora explicitly.

Possible but harder — less training data available. We'd source corpora and annotate more heavily. Timeline and cost go up accordingly.

For defect detection: 2,000-5,000 labelled images per defect class, typically. We help design the collection strategy.

Factory lines: edge (Jetson). Retail analytics: cloud is fine. Privacy-sensitive (face): edge, always.

Lighting is 80% of production CV. We spec lighting rigs as part of the solution — you'll never get a good model on bad pixels.

Both. Batch for candidate set generation (nightly). Real-time for ranking (per request). Hybrid is the production norm.

Handled via content-based fallback for new users / items, popularity for no-context cases, and explicit onboarding preferences where possible.

Depends on the problem. Demand forecasting: 8-15% MAPE is excellent. Churn: 70-85% AUC. We set realistic targets based on problem complexity.

Helpful, not required. We'll work with your analyst / ops lead. For ongoing retraining, we can operate the pipeline remotely.

Models are always somewhat wrong. We ship with confidence intervals, degrade-gracefully fallbacks, and human-override paths for high-stakes decisions.

Depends on use case and sensitivity. Non-sensitive: OpenAI is still the benchmark. Sensitive: Claude or on-prem Llama 3. Cost at scale: Gemini or Mistral. We benchmark on your data.

Cache common responses, use smaller models for classification, fall back to bigger only when needed. Cost is roughly 30-40% of naive usage if done right.

For regulated industries, we deploy Llama 3 / Mistral on your AWS / Azure. No data leaves your cloud. Slightly lower accuracy, zero compliance risk.

TanStack Query for server state, Zustand or React Context for client state. Redux only when legacy code requires it.

App Router for all new projects. Pages Router is perfectly fine for existing codebases — we don’t force rewrites.

Fastify for simple APIs, Nest for large teams needing opinionated structure, Express for legacy maintenance only.

Django for full-stack apps with admin, auth, and ORM built-in. FastAPI for APIs-only, especially AI-adjacent.

Containers for steady load, Lambda for spikes and glue. Most production systems use both. We design the split on actual usage.

GraphQL when many clients (web, mobile, partner) need different slices of the same data. REST for public APIs and simple CRUD.

Yes. You interview them on your process. We filter for seniority and fit first. You make the final call.

Replace within 5 days, no charge. The 3-day paid trial exists precisely so misfits don't become 3-month mistakes.

We do. Everything. You pay one monthly invoice. The developer is on our payroll, using our laptop, with our HR team.

IST gives 4h overlap with UK (2pm-6pm IST / 9am-1pm BST), 4h with US East (5pm-9pm IST / 7am-11am EST), 5h with AUS (9am-2pm IST / 1pm-6pm AEST).

Yes. Clear IP-assignment in every contract. Your code, your product, your IP — always.

Yes. 30-day notice to add or remove. We don't lock you into annual contracts.

Target 15-30 meetings per caller per month for B2B SaaS. Higher for lower ACV, lower for enterprise.

No agency worth hiring guarantees meetings in month one. We commit to dial volumes, script rigour, and weekly iteration. Meetings follow.

You do. All call recordings, CRM data, and lists live in your systems. We access them on your behalf.

Tamil, Hindi, Telugu, Kannada, Malayalam, Marathi, Bengali, English as standard. Others on request.

Yes. We integrate with Zoho, HubSpot, Salesforce, or your custom CRM. All calls logged in real-time.

Random call sampling (5-10% of all calls scored weekly), scorecard feedback, monthly coaching sessions.

Official only. Unofficial automation gets numbers banned, costs trust, and breaches Meta policy.

Relevance + spacing + opt-out respect. Every message advances the conversation. Every lead can opt out in one tap.

You do. All call logs, chat transcripts, and email threads sit in your CRM.

Product knowledge, script training, tone-of-voice guide, shadow-mode before live, weekly calibration calls.

We replace within 5 days of any exit. Parallel shadowing means transitions are seamless to you.

Sample recording reviews weekly, scorecards, monthly calibration. You get a monthly QA summary.

Standard 9-6 IST, extended 8am-8pm, or 24/7. Your hours, our coverage plan.

Yes. We’ve shipped healthcare products that passed regulator scrutiny. That said, we pair with domain legal counsel rather than replacing them.

We’ve integrated with the main healthcare platforms and legacy systems. API, file-based, or database-level integrations handled.

Tamil, Hindi, and English are standard. Other regional languages on request — always human-reviewed, not raw machine translation.

We prepare, validate, and support one-click filing. Final submission is the user’s responsibility — always under your DSC or EVC.

We queue and retry. For time-sensitive invoices, we give the user visibility and alternative options.

96%+ on common categories using our ML classifier, with human-override. Ambiguous items flagged for review.

Multiple, almost always. Primary + fallback routing gives 2-5% higher success rates at the cost of one extra integration.

OTP confirmation before order creation. Pincode-level COD toggles. Risk-scored customers.

Stripe or Razorpay International for INR-to-INR export. RBI LRS compliance for outbound. We handle both.

BSPs are official partners of Cloud API. Gupshup, Wati, Interakt — all fine. We pick based on your volume and specific needs.

Never. They violate Meta’s ToS and your number will eventually be banned. We only build on official API.

Standard templates: 1-24 hours. Complex ones (media, rich): 1-3 days. We write them to maximise first-submission approval.

Fine for translation and basic TTS. Not fine for domain-specific Tamil — medical, legal, financial. We fine-tune on your domain.

Yes. Explicitly trained on mixed-script corpora. "vanga saapdalam" is parsed correctly.

Yes. Printed (95%+ accuracy) and handwritten (70-85% depending on quality).