Getting started
I don't have a full spec yet. Can we still start?
Yes — that's the default, not the exception. Every MVP kicks off with a 2-day brief workshop where we turn your bullet points, screenshots, and Loom recordings into a working spec. You don't need to write requirements. That's our job.
Getting started
What's included in an MVP, and how is it priced?
Most MVPs take 4–8 weeks. After a short scoping call we lock the scope, price, and launch date in writing, so the quote is fixed before we start, and if we miss the agreed date, the extra time is on us. The work: two senior engineers and a senior QA engineer, directing AI agents designed and tuned in-house. The engineers own architecture, security, code review, and production deployment; QA tests every release end to end; the agents carry out much of the coding, test automation, and documentation. Includes AI tooling, cloud setup, senior architecture + security review, and 30-day post-launch support. Scope-wise: one product, one user role, the core happy-path workflow, and auth + billing wiring if you need it. Whatever stretches that — second user role, complex integrations, mobile app in parallel — moves to a follow-on phase with its own quote.
Getting started
Why not just build the MVP with AI tools myself?
Honest answer: you can. Lovable, Cursor, v0, Claude Code — they'll get you to a working demo fast. The cost shows up in the next six months. AI tools alone don't design architecture that scales past the demo. They don't harden auth, security, or edge cases. They don't wire CI/CD, observability, or alerting. They can't continue building once the prototype is up. We use the same AI agents you would, but with two senior engineers steering them and reviewing every commit, a senior QA engineer testing the product end to end, and production code shipped into your repo and your cloud. You're paying for what AI tools alone can't do: the senior judgment that turns a demo into software you can put real customers on.
Getting started
What if I need something bigger than the MVP?
The MVP is the productized entry — one product, one user role, the core happy-path workflow. For deeper specialized work (production AI agents at scale, voice pipelines, multi-tenant platforms, custom-trained models, multi-year embedded teams), we scope custom engagements with their own timeline and dedicated-engineer counts. Same $1,800/month per engineer rate, same senior review + deploy process, just sized to what you actually need. Talk to us — we'll quote in 24 hours.
Getting started
How are you this much cheaper than US agencies?
Three reasons. (1) Our engineering center is in Lahore, Pakistan — senior talent at $1,800/month per engineer, not Bay Area loaded cost. (2) AI agents handle the bulk of code production; engineers steer scope, architecture, security, review, and deploy. We ship more per engineer than a traditional team. (3) No agency overhead — no salespeople, no account managers, no Toptal-style commission layer. You pay for engineering capacity, not the agency machinery around it.
Getting started
How do payments work?
For an MVP: 50% on kickoff, 50% on production deploy. Retainer engagements invoice monthly at the start of each cycle. On-demand work is billed at $15/hour for the hours used. USD, wire or Stripe. We don't take equity or success fees.
Getting started
Do you take small projects or tight budgets?
Yes. There's no minimum project size. Small builds, updates, and fixes run on demand at $15/hour (limited offer for new clients; regular rate $20/hour), and you pay only for the hours used. Tell us your budget and what you need, and we'll suggest the option that fits within 24 hours.
Getting started
What tech stack do you use?
Frontend: Next.js / React with Tailwind. Backend: Node or Python, whichever fits the workload. Data: Postgres or Supabase by default; Mongo or Redis when the use case calls for it. AI: Claude, OpenAI, or open-source models — your choice. Infra: deployed into your cloud (AWS, GCP, Vercel). We don't lock you into our stack — you'll be able to hire any senior engineer to extend it.
Getting started
Do we need our own cloud account?
Yes, and that's intentional. We deploy into your AWS, GCP, or Vercel account from day one — never to our own infra. The codebase, the database, and the bill all live with you. If you don't have one set up yet, we'll spin it up with you in the first week.
Getting started
What if I need changes mid-sprint?
Small adjustments to the working scope are part of the price — we don't nickel-and-dime on copy tweaks or layout shifts. Net-new features or scope additions get a quick written change order with a delta in days and dollars before any work starts. No surprise invoices.
Getting started
How often do we meet?
One 30-minute demo call every Friday with a working build to look at. Async daily updates in a shared Slack or Discord channel — scan in the morning, reply when you have time. Sync calls only when there's a real blocker. We don't pad the calendar with status meetings.
Getting started
Who owns the code and IP?
You do, fully, from day one. The code lives in your repo, the cloud account is yours, and there's no licensing back to Setubo. We don't retain rights to your codebase, your data, or anything we build together.
Getting started
Will you sign an NDA?
Yes — standard mutual NDA before any technical discovery. Use your template or use ours, doesn't matter. We've never had a deal stall on legal.
Getting started
What happens after the 30-day support window?
Two paths. Most clients move into a monthly retainer — same team, ongoing iteration, predictable cost. A few take the keys and run with it themselves; we'll do a handoff session, document anything custom, and stay reachable for one-off questions. No lock-in either way.
Getting started
What if it's not working out?
Two-way exit. You can wrap the engagement at any sprint boundary; we'll return the in-progress code, document the state, and refund unbilled time. If we sense bad fit on our side — wrong domain, wrong stage, scope you'd be better off DIYing — we'll surface it before week 2, not after the invoice.
