Services
The Team Behind Our Own Product, Available for Yours
We are not a consultancy that has never run anything in production. We build and operate C-Band MailBox, which means payments, mail handling, operator payouts and AI inference costs land on us every day. We take on a limited amount of outside work from the same team.
What We Do
Four Kinds of Work
Product engineering
Building and running the actual thing: web and mobile front ends, an API behind them, payments, notifications, admin tooling and the unglamorous operational work that decides whether a product survives its first hundred users.
- Modern JavaScript front ends
- API and database back ends
- Stripe payments, payouts and invoicing
- Native iOS delivery
- Admin and internal operations tooling
- Multi-language products, including French for Canada
Applied AI
Putting a model into a product that already has users, which is a different job from building a demo. We have done this in our own platform and can do it in yours.
- Document and image understanding with vision models
- Summarization and extraction from scans and PDFs
- Prompt design kept in source and reviewed like code
- Background jobs that fail without taking the app down
- Cost controls so inference spend cannot run away
- Opt-in and privacy handling for user-facing AI
Integration work
Connecting a system to the other systems it has to live with. Most of the difficulty is in the edges, the retries and the reconciliation, not in the happy path.
- REST and GraphQL API integration
- Payment and payout provider integration
- Transactional email and notification delivery
- Object storage and file pipelines
- Webhooks with retry and idempotency handling
- Data migration between systems
Technical assessment
A straight answer on whether the thing you are considering is worth building, and what it will actually cost to run. Sometimes the honest answer is that a model is not the right tool.
- Feasibility review before you commit budget
- Build versus buy on AI features
- Running-cost modelling for inference and infrastructure
- Architecture review of an existing system
- Scoping a first release that can ship
The Proof
What We Have Actually Shipped
The clearest thing we can show you is our own platform. C-Band MailBox is a two-sided system: renters manage mail that arrives at a physical location, and the operators of those locations run their business on the same software, with their own plans, their own pricing and their own tax handling.
That means we have shipped and now maintain payment flows and operator payouts, per-operator public websites with DNS provisioned automatically, recorded terms acceptance that can be exported as proof, a document pipeline moving scans through object storage, transactional notifications, four languages, and a native iOS app, in four languages.
On the AI side, the summarization feature runs a vision model on rented inference. It is opt-in per user, runs in the background without being able to break the request that triggered it, falls back through several document sources, and refuses to regenerate a summary it has already produced so a stray retry cannot cost money. That whole design is written up on the AI page.
How Engagements Run
No Surprises at the End
Conversation
A technical conversation about what you are trying to do and why. If we are not the right team for it, this is where we say so.
Scope
We write down what will be built, what it will run on, and what it will cost to operate. You approve that before any code exists.
Build
Short cycles with something you can look at and use at the end of each one. No long silences followed by a reveal.
Handover or run it
We hand the system over with its documentation, or keep operating it for you. Both are fine, but we agree which one up front.
Get Started
Tell Us What You Are Building
Send us the problem rather than the specification. We will tell you what we would do about it, what it would take, and whether we are the right people for it.
Talk to us