Independent software lab
We design and ship applications, AI Models, and games.
GASPEROHLAB takes hard problems from a single question all the way to production — engineering software people actually use, not slide decks.
- Based
- Toronto, Canada
- Founded
- 2025
- Disciplines
- Games · Apps · AI · Systems
- Model
- Independent, no investors
01/What we build
Four disciplines, one team
Every project starts as a question. These are the forms the answers take — heaviest first.
01/Product
Applications9 shipped
Web and mobile products built with intent. Fast, accessible and genuinely pleasant to use — the bulk of what the lab ships.
02/Infrastructure
Systems & tooling2 shipped
The engines, analysers and command-line tools underneath everything above — and, lately, two that ship under their own names because the hard part was never the interface.
03/Interactive
Games2 shipped
Playable systems, built the same way as everything else: the simulation has to be true before any of it is fun. Mechanics first, polish always.
04/Intelligence
AI models1 shipped
Models fine-tuned on one industry's documents, and the infrastructure to run them without renting an AI vendor's. One entry, and the largest thing here — the platform every construction tool below is built on.
02/Selected work
What we've shipped
Products in production today, with more taking shape in the lab. Open any project for the full story.
03/In production
“We needed our own model on our own infrastructure — no AI vendor in the path, reading construction documents properly rather than generically. The lab built the engine and the five programs on top of it, and they're in production.”
Our work isn't a portfolio of concepts. The Marapone suite — five construction programs on a model we built and run ourselves — is built here and open in contractors' browsers today.
04/The numbers
Proof, not promises
What the lab has actually put into production, and the guarantees that come with it — counted from the archive, not estimated.
6
Products released into production
7
More being built in the lab right now
0
Bytes of client data that leave the building
100%
Source handed over — no black boxes
05/How we work
From a question to production
No two projects look the same, but they all move through the same four gates. Most ideas don't make it past the second — and that is exactly the point.
- Step 01
Ask
We start with a question worth answering, not a spec. If we can't say why it matters, we don't start.
- Step 02
Prototype
The smallest real version we can build, put in front of the actual problem — early, while walking away is still cheap.
- Step 03
Pressure-test
We break it on purpose. What survives contact with real use moves forward; what doesn't gets cut, cleanly.
- Step 04
Ship & own
What earns it goes to production — engineered, accessible, and still ours to maintain long after launch.
06/Ways to work with us
Three ways in
Most engagements start small and earn their way forward. Pick the one that fits where you are — or just tell us the problem and we'll say which makes sense.
Prototype sprint
Weeks, not months. We build the smallest real version of your idea and put it in front of the problem — so you learn whether it's worth going further before committing to a roadmap.
- A working prototype
- An honest read on feasibility
- A clear go / no-go
Build to production
The prototype survives contact with reality, so we take it the rest of the way — engineered, accessible and shipped. You own the code and the decisions behind it.
- Production-grade build
- Full source, no black boxes
- Ongoing ownership
Private AI deployment
A model fine-tuned on your industry's documents, running on infrastructure with no third-party AI vendor in the path — the same way we build and run MaraponeAI. Into your own tenant where the contract demands it.
- Trained on your domain
- No third-party AI vendor
- Your tenant, if required
07/From the lab
The thinking behind the work
Engineering write-ups and design notes on how the products are built — and why some of them were cut.
08/Before you ask
How working with the lab works
The short version of the questions we get most. Anything not here, just ask us directly.
Hard, interesting problems across games, applications, AI models and the systems that hold them together. If it can be prototyped, pressure-tested and shipped — and we can say why it matters — it's a fit. If it's a slide deck looking for a rubber stamp, it isn't.
With a conversation and a prototype, not a 40-page statement of work. We make the smallest real version we can and put it in front of the problem early, so we both learn whether it's worth building before anyone's committed to a long roadmap.
Yes. Commissioned work ships with its source and the reasoning behind it — no black boxes you can't open, no dependencies you can't replace, and nothing you'd have to buy back from us later. Our own products are priced however they're priced; what we build for you is yours.
It's a core competency. MaraponeAI is our own model on our own infrastructure — no commercial AI API anywhere in the path, so nothing gets forwarded to OpenAI, Anthropic or Google. Where a contract requires that documents never leave the client's own environment, we deploy into their tenant instead. If data sensitivity or cost rules out a third-party hosted API, this is exactly the kind of thing we do.
A working prototype is usually weeks, not months — that's the whole point of moving as a lab. Production timelines depend on scope, but we keep the loop between idea and shippable software as short as we honestly can.
It depends on what you're building, so we won't pretend a number here means anything. Tell us about the problem and we'll come back with an honest scope — including whether we think it's worth doing at all.
Taking one new project for Q4 2026
Have something worth building?
A game, an app, an AI model, or a problem that doesn't have a name yet — tell us about it and we'll build the prototype.