Service · MVP / Web App
Turn the idea into a working product.
A real, usable first version you can put in front of customers or investors. Production-grade, not a throwaway prototype — and built to grow into the full platform when you need it.
From £12,500
What you get
Working software, not a mockup.
Every MVP ships with the foundations that let it survive real customers — auth, data, security, observability — built in from the start.
Working core flows
The key user journey — whatever makes the product useful — built and tested on real data. Not a mockup, not a proof of concept. A thing that works.
Auth and data model
Sign-up, sessions, password reset, TOTP. Per-row security enforced at the database. A schema designed for the next two years, not just the demo.
The key feature
The one capability that makes the product worth building. Built as a first-class part of the system, not glued on at the end.
Deployed to your cloud
Your Vercel account, your Supabase project, your domains, your keys. We deploy and walk off. No agency-held infrastructure, no lock-in.
Search that works
Fuzzy keyword + trigram search inside Postgres — typos don't break results. No Algolia bill, no third service to keep alive. Upgrade to semantic search later.
File uploads and storage
Magic-byte validation before anything lands in storage. Size limits, filename sanitisation, auto-rollback on failure. The basics done right.
Observability from day one
Structured logs, error reporting wired to Sentry, and latency dashboards before launch — not retrofitted after the first incident.
CI/CD and preview deploys
Every pull request gets a live URL. Type-check and lint gate merges. Deploys are a button. Rollback is a button.
Why it’s real software
Built on the same foundation as the full platform.
canarlo-core
The engine behind every build
Config in, full production application out — every MVP already ships with this foundation.
BuildauthCMSpaymentssearchstorageAIchatagentsRAGsummariesEnforce8 security patternsauto-migrationstyped SDKWhat grows with you
The Production Platform path
MVP is layer one. You add capabilities on the same codebase — never rewrite the foundation.
Add laterbillingmulti-tenancysemantic searchAI featuresNeverarchitecture rewritenew codebaserebuild costs
How it works
From first call to working product.
Discovery
One scoping call, then a written brief. The load-bearing pieces, the failure modes, what you've already tried. Two weeks. No workshops, no decks.
Architecture
Schema, API surface, build plan — on the page. Data model and security boundary named. Two weeks. You sign off before a line is written.
Build
Eight to twelve weeks. Weekly preview URL, readable diffs, every integration audited from day one. You can read what shipped and why.
Ship
Deployed to your cloud, keys transferred, runbook delivered. A handover doc that names the failure mode and the on-call step. Not a goodbye email.
Ready to build the thing that matters, not a throwaway prototype?
Questions
What founders ask before they commit.
What's the difference between an MVP and a prototype?
A prototype proves the concept; an MVP proves the business. What we build is production-grade software — real auth, real data model, real security — not a throwaway you rebuild once it gets traction. The architecture is designed to grow: the same codebase can become your full platform later, adding payments, search, admin, and more without a rewrite.
How long does an MVP take to build?
Two weeks discovery, two weeks architecture, eight to twelve weeks build. The exact timeline depends on the scope defined in the architecture phase — not guessed at kickoff. Every phase ends with a concrete artefact: a brief, a schema, a deployment.
Who owns the code?
You do. Your repo, your Vercel account, your Supabase project, your keys. No Canarlo SaaS in the loop. The handover doc names every environment variable, every cron job, and every integration. You can take it in-house tomorrow and the system keeps running.
What if the scope isn't clear yet?
Start with discovery. Two weeks, fixed-fee — one scoping call, a written brief, the load-bearing pieces named and the failure modes mapped. You leave with enough to commit to the build, or to take the brief to another studio. Either is fine.
Can the MVP grow into a full product later?
Yes — deliberately. The architecture is chosen so the foundation doesn't get thrown away when you add billing, AI, semantic search, or a full admin. Payments, job queues, additional auth flows, integrations: all can be layered in. You don't pay to rebuild the foundation.
Start here
One call, honest answer.
Tell us what you’re building and where you are in the process. If it fits, we’ll say so on the call. If not, we’ll point you somewhere useful.