Pineau Labs · Charlottetown, PEI
Your idea. Working software.
2 to 4 weeks.
I'm Brad Pineau. I've spent 25 years shipping software, and I build working prototypes for a fixed price, typically $5,000, with the same agentic AI tools I run in production every day. If the idea works, the same codebase grows into your MVP. Nothing gets thrown away.
Brad Pineau
Engineer & founder. Former CTO. 4 software patents.
25 years shipping production software
Former CTO who acquired and scaled a tech company
4 software patents
A decade of HIPAA healthcare systems
Start small, scale what works
One path from idea to product.
Most clients start with a prototype. Every step is fixed-price and time-boxed, and each one builds on the last. Prices shown are typical for a standard scope; something simpler costs less.
Prototype Sprint
Your idea working on a live URL. Real auth, real data, and two or three core flows you can put in front of investors and users.
Sprint detailsMVP Sprint
The same codebase, hardened for first paying users. Stripe payments, monitoring, admin tools, and a security baseline.
Sprint detailsKeep shipping
Senior engineering leadership on retainer. Keep iterating after launch without funding a full-time hire. Cancel anytime.
How it worksAlready have a product? The AI Integration Sprint adds real AI features to it. Selling to enterprise buyers? Start with the 1-week Security Sprint. Or browse the full catalog.
Why start here
The prototype isn't a throwaway.
It's phase one.
Agencies will quote you $25K and three months to find out if your idea works. Vibe-coding tools will hand you something in a weekend that collapses the first time a real user touches it. Both leave you starting over.
My prototypes are built like real software from day one: real authentication, a real database schema, a standard stack any engineer can maintain. The repo lives in your GitHub from the start. When you're ready for paying users, we don't rebuild. We keep going on the same codebase.
Agentic AI is what makes the price and timeline possible. Agents handle the parallel build work; I make every architecture call, and I review and QA every update and build before it ships. It's the same workflow I use on my own products every day.
Of agent output reviewed and QA'd by me before it ships.
The GitHub repo is yours, on a standard stack with no lock-in
Frontier models cross-verifying every high-stakes call
AI tools shipped on Fitly, my own product, with this exact playbook
Retainers or obligations after handoff. Continue only if it's working
Proof
Recent builds
Fitly AI
My own product and the flagship proof: a chat-first fitness app where one agent runs a user's whole fitness life, logging meals from a photo, building workouts, and tracking macros through 52 tools, all wrapped in a ten-layer orchestration engine that makes the AI reliable enough to trust with real data. Same playbook, same stack, same workflow I bring to client work.
AI groundwork and modernization for a multi-tenant healthcare platform serving hospice care providers.
A cause-driven payments platform where supporters buy prepaid cards and every card funds the charity behind it.
A decade of clinical software at Timeless Medical Systems.
How it works
From first call to handoff.
Email me the idea
Use the form or write brad@pineaulabs.com. Rough is fine: notes, a sketch, a voice memo. I'll get back to you personally with which sprint fits and what it'd take.
Fixed-price scope
A few emails later you have a written scope and a fixed quote. Price and timeline are agreed before any work starts.
Watch it grow
Agents build in parallel while I review and QA every update. You get regular progress updates over email, so you see real progress without a meeting on your calendar.
It's yours
Live URL, the GitHub repo, a written what-next plan, and a live demo over Zoom. No retainer, no lock-in.
Is this for you?
Who this is for, and who it's not.
Best for
- Founders with a clear product idea who need it built
- Teams validating a workflow before a big build
- A business with one painful manual process to automate
- An existing product that needs real AI features, not a chatbot
Not the right fit
- Vague research projects with no target outcome
- Enterprise, procurement-heavy builds with long approval chains
- Apps that need a big design phase before any code
- Anyone shopping purely for the cheapest quote
What clients say
"Brad helped develop quite a few projects, from coding to app and web development, and can pretty much build anything that's possible to build. I've known Brad for more than 15 years, and he's one of the most skilled and reliable people I know. Brad delivered many projects for me with flying colors."
Patrick Thona · HospicePro
Common questions
What founders usually ask first.
Who owns the code?
You do. The repo lives in your GitHub account from day one, on a standard stack (TypeScript, Next.js or Angular, Postgres) that any engineer can maintain.
I'm not technical. Is that a problem?
Not at all; most of my clients aren't. Bring the idea in whatever form it exists: notes, sketches, a voice memo. Progress updates are in plain English.
Is it really fixed price?
Yes. After a short email exchange you get a written scope and a fixed quote. If scope changes mid-sprint, you approve the new number before I build anything extra.
How do payments work?
A deposit reserves your build window; the balance is due on delivery. The exact split and the price are in your written quote before anything starts.
What happens after the prototype?
Whatever you want. Take the repo and run, continue into the MVP sprint on the same codebase, or keep me on as a fractional CTO. No obligation either way.
Do we need to get on a call?
No. Scoping, progress updates, and handoff all run over email, so the project never waits on a meeting. If you'd rather talk something through on a Zoom, that's always an option.
Will you sign an NDA?
Happy to. I'll sign your NDA before you share anything sensitive, and I keep client work confidential by default. The example builds on this site are deliberately anonymized for exactly that reason.
Do you train AI on my code or data?
No. Your code and data are never used to train AI models. The tools I work with run under commercial terms that exclude your content from training, and your code stays in your own GitHub and cloud accounts.
Email me what you're working on.
I'll get back to you personally with which sprint fits, or a custom scope if a different shape is better. Everything runs over email unless you'd rather jump on a Zoom. No pitch, no sales calls.
Get in TouchBased in Charlottetown, Prince Edward Island. Working with Island businesses and clients worldwide.