The Chilli Index
A searchable database of hot sauces with AI-powered taste profiling and recommendation. Built in a series of evenings using AI as the development partner.
SHIPPEDthechilliindex.com
Expansion · 4 weeks · From £2,500
Got an AI product idea? Prove it in 4 weeks.
A fixed-scope sprint to scope, design, and build a working prototype of your AI-powered product or feature. Validate the concept before committing to a full build.
4week sprint
£2,500+fixed scope
1working prototype
What this is for
You've got an AI product idea - a tool, an assessment, a personalised generator, a client portal. Something that uses AI in a specific way for a specific audience. The question isn't “should we build it” - it's “what should we build, and will it actually work?”
A Product Sprint answers that with a working prototype in four weeks, fixed scope - a real thing you can put in front of real users to validate the idea before committing to a full build.
The four weeks
Each week has a defined output. By the end of week 4, you have a working prototype, a technical specification, and a clear view of what a full build would look like.
Discovery and scoping
5 working daysOutput: a written scope document and technical architecture sketch.
- Define the prototype's core user journey
- Identify the technical stack and integrations needed
- Confirm what's in scope and what isn't
- Surface assumptions and risks early
Prototype build
5 working daysOutput: a working build accessible at a staging URL.
- Build core functionality end-to-end
- Wire up AI integrations (Claude API typically)
- Implement the user journey defined in week 1
- Stand up staging environment for review
Testing and iteration
5 working daysOutput: refined prototype tested against real users (or your test scenarios).
- Put the prototype in front of 3-5 sample users (or run your test scenarios)
- Identify what works and what needs adjusting
- Iterate on the highest-impact issues
- Document what we learned
Handover and GTM
5 working daysOutput: documentation, go-to-market recommendations, and a decision document.
- Full technical documentation handed over
- Go-to-market recommendations based on testing
- Build estimate for a full production version
- Go/no-go decision document - proceed, pivot, or park
How pricing works
Sprints start at £2,500 and scale based on the complexity of what's being built. Most fall in the £2,500-7,500 range. We agree the price upfront based on the week 1 scoping conversation - no scope creep, no surprise invoices.
Typical scope ranges
Simple prototype
Single-purpose tool, Claude API + Next.js, minimal user system
From £2,500
Standard prototype
Multi-step user journey, auth, database, integrations
£4,500-7,500
Complex prototype
Multiple AI workflows, payments, third-party APIs, admin tooling
From £7,500
The 30-minute discovery call confirms which range your idea sits in. If your idea is bigger than a Sprint can sensibly cover, we'll say so on the call - and recommend the right next step.
Who this is for
- You have an AI product idea but aren't sure if it'll work in practice
- You need a working prototype to show investors, leadership, or pilot users
- Your existing business has an AI feature opportunity you want to test before fully committing
- You want a fixed-scope, fixed-price engagement with a real deliverable at the end
Who this isn't for
- You need a production-ready product launched at the end - Sprint produces prototypes, not finished apps
- You're not sure what you want built - Sprint needs a defined idea to scope against
- You want ongoing iteration without a defined endpoint - that's a retainer arrangement, not a sprint
What's possible
Things we've built. Things you could build.
Some of these are our own AI-built projects. Some are sprint candidates we've scoped. None are real-client case studies yet - we'll swap these for actual sprint clients as we deliver.
Dadhmor
A mobile app for new dads, with AI-generated daily prompts and parenting advice tailored to baby age. Built in a couple of days using AI tooling.
SHIPPEDdadhmor.app
Mindful Thought of the Day API
An API delivering a daily mindfulness prompt to corporate intranets, digital signage, screensavers, and email signatures. Scoped from an existing meditation content library.
SCOPEDready to build
Client onboarding assistant
A guided form that asks new clients smart, context-aware questions, then produces a structured onboarding brief for the agency team.
SCOPEDready to build
Personalised assessment tool
A quiz-style assessment that produces a personalised AI-generated report for each user - usable as a lead magnet or paid product.
SCOPEDready to build
Something specific to you
If you've got an AI product idea - a tool, a service, an internal capability, a customer-facing feature - describe it on the call. We'll tell you if it's sprint-shaped.
BOOK A CALLlet's scope it
Sprint products usually run on Next.js + Supabase + Claude API. Other stacks possible - ask on the call.
This is usually the point people book a call.
Book a 15-min call →15 minutes. Honest answer, even if it's “not yet” or “not us.”
Common questions
Things worth knowing before you book.
If something isn't covered, ask on the call. Honest answers, no sales pitch.
What's the difference between a "prototype" and a real product?
A prototype proves the idea works. It runs on staging, handles the core user journey, integrates with the AI properly, and lets real users (or you) test the experience. It's not production-ready - it doesn't have full error handling, scaling, monitoring, edge case coverage, or polish on every screen.
At the end of week 4, you'll know whether the idea is worth building properly. If yes, we estimate a full production build separately. If no, you've spent £2,500 to learn that - much cheaper than building the wrong thing for £50,000.
How is the price decided?
Week 1 is the scoping conversation. We define what's in scope, what's out, what stack, what integrations. By the end of week 1, we agree the fixed price for the remaining three weeks. If the work is bigger than what fits in £2,500, we'll quote it accurately - or recommend scaling the scope back so it fits the entry tier.
The £2,500 starting point covers the scoping week even if you decide not to proceed with the build. So you're never committing to the full sprint blindly.
What stack do you use?
Most sprints run on Next.js (App Router) + Supabase + Claude API, deployed on Vercel. This stack is fast to build with, scales well, and handles AI integration cleanly. It's also what most clients can hand off to their own developers afterwards.
If your idea needs a different stack (mobile app, specific integrations, existing platform), we'll discuss it on the call. We're not religious about the tech - but the default exists because it works.
Who owns the code at the end?
You do. The code lives in your repository (GitHub, GitLab, wherever). You get full access from week 1. The technical architecture, deployment config, environment variables, everything is documented for handover.
There's no platform lock-in. If you want to take the code and rebuild from scratch with your own developers, that's a valid path - and the documentation supports that.
What happens after the sprint?
Three common paths. First, you take the prototype and your in-house developers build it out properly - that's the cheapest route if you have the team. Second, we estimate and deliver a full production build as a separate engagement. Third, you put the prototype in front of users to validate further before deciding.
About a quarter of the sprints we'd expect to do end with "the idea didn't work and we're glad we found out for £2,500 rather than £50,000." That's a successful outcome too.
Do I need to be technical?
No. You need a clear idea of what you want to test and the ability to make decisions. We handle the technical work. We translate technical trade-offs into language that makes the decisions easy - e.g. "this would take an extra week, or we can ship it without this feature and add it later."
If you ARE technical and want to be involved in the build, that works too. Some clients want to learn from the process. The format adapts.
Do I need an AI Blueprint first?
Not necessarily. A Sprint can be a standalone engagement focused entirely on the new product. A Blueprint helps if you don't yet have a clear sense of which AI product to build - its opportunity map identifies candidates. If you already know what you want to build, go straight to the Sprint.
If you have both, the prototype can reuse the brand voice, audience personas, and other context from the Blueprint - which makes the build faster and more on-brand.
Got an idea? Let's pressure-test it.
15 minutes to walk through what you want to build. Honest answer on whether a sprint will validate it - or whether you need something different.