Web app, mobile app, or both — deployed in your own cloud account, with real accounts, real data and the AI features working end to end.
Chat, copilot, document workflow, RAG search over your own content or an agent that calls your tools — whichever the problem actually needs, with evaluations so you can see quality before your users do.
The ingestion, cleaning and indexing behind the AI feature, running on a schedule rather than in a notebook someone runs by hand.
Your GitHub organisation, your cloud account, commit by commit from week one. Infrastructure as code, so it can be rebuilt without us.
Rate limits, fallbacks when a model is slow or down, logging of prompts and outputs, and cost tracking per feature.
Architecture notes, runbooks and a walkthrough with your engineers. Plus a 1-month warranty on what we built.
We pin down the one job the MVP must do, the data it needs, and what is explicitly out of scope. You approve screens and the model plan before code is written.
The product surface and the AI layer, built together. First clickable demo in week 3 — you can use it, not just watch it.
Your data, your integrations, evaluations on the AI output, guardrails, and the unglamorous parts: auth, roles, error states, cost controls.
Deploy into your cloud, walk your team through the code, fix what the walkthrough surfaces, and start the warranty month.
We quote after scoping, not before. The number doesn't move unless you change the scope, and what's excluded is written down next to what's included.
You pay against delivered phases. The final payment is the smallest, and it comes after handover.
If something is slipping, you see it in week two — not in week nine when the deadline is gone. Demos are recorded, so whoever missed the call can still watch.
Your repo, your cloud, from the first commit. Full rights transfer on final payment, and since payments are milestone-linked, that last one is small.
Signed before you share anything confidential. Scope, milestones and acceptance criteria live in the statement of work.
Infrastructure as code and documented handover mean another team can pick it up. There's no platform of ours to stay locked into.
We're on India Standard Time (UTC+5:30). That's a comfortable overlap with the UK, Europe, the UAE and Australia, and an early-morning window with the US East Coast. For the US West Coast, live overlap is short: we run on written hand-offs, and your morning starts with an update from our day.
A core team of 7 to 16 people, with a lead and a project manager on your work — not a rotating bench. You meet the engineers on the project and they stay on it.
We deploy into your AWS, GCP or Azure account, with role-based access. For sensitive work we run open-weight models inside your cloud instead of calling an external API.
If an AI feature should be a database query, we'll say so before you pay for it. A scoping call where we talk you out of something is a good outcome for both of us.
We're not a 500-person outsourcer with an account manager between you and the work, and we're not the cheapest quote you'll get. We are senior people who ship, at Indian rates rather than US agency rates.
Scope drives the price, so we quote after a short call rather than putting a made-up range on this page. Here's exactly how the money works, so nothing surprises you later.
If your budget is fixed, tell us the number on the call. We'll say honestly whether it fits a 60-day MVP or not.
The scope is fixed in the first ten days precisely so that doesn't happen, and the weekly demo shows you the truth early. If we're the reason a milestone slips, we finish it — the bid is fixed. If the scope changes mid-flight, we re-quote the difference in writing before doing the work.
The code is written in your repository and deployed to your cloud from the start, so you always have it. Full rights to the bespoke work transfer on final payment, as set out in our terms. Pre-existing tools and libraries we bring are licensed to you as part of the deliverable.
No. We deploy into your own cloud account, and the commercial model APIs we use don't train on API data by default. If your policy doesn't allow external APIs at all, we run open-weight models inside your infrastructure instead.
Role-based access, audit logging, encryption in transit and at rest, and deployment into your own cloud region — so residency is your choice. We're ISO 9001:2015 certified, which covers how we run projects, not security; the security controls above are what actually protect your data. We work to GDPR, UK GDPR, PIPEDA, Australia's Privacy Act and India's DPDP depending on where you operate.
Yes. A lead, a project manager and the engineers building your product stay with it from scoping to handover. You'll meet them on the scoping call, not after signing.
You have a 1-month warranty on what we built. Most clients then move to a monthly retainer for feature work, dependency and security updates and monitoring. Some take the code in-house — that's a fine outcome, and the documentation is written for it.
Thirty minutes. You describe the problem and who it's for; we ask about data, integrations and constraints, and tell you what we'd cut to fit 60 days. You get a written scope and a fixed quote within one business day.
Yes. We build against your API contracts, work in your repository conventions, and review with your engineers. If you have a design team, send Figma and we'll build to it.
Whichever fits cost, latency and your data rules: OpenAI, Anthropic, Google, or open-weight models such as Llama, Qwen or Mistral running in your own cloud. We're not tied to any provider.
Yes — mutual NDA, then your MSA and SOW or ours. US, UK and EU-law contracting is routine for us.
Start with the scoping call anyway. If a report or a database query solves your problem, we'll tell you, and you'll have lost half an hour rather than a quarter.
Yes — iOS and Android from one codebase in Flutter or React Native, alongside the web app and the AI layer, by the same team.
A 30-minute call, then a written scope and a fixed quote within one business day. If 60 days isn't realistic for what you need, we'll say so on the call.