You'll always know where things stand.
No black box, no ceremony for its own sake. We trust professionals to be professional, and we work in the open so you can see it.
Reality is the source of truth
We take the Agile Manifesto at face value, which is exactly why we don't run much ceremony. Standups, sprints and story points are tools. We use them when they earn their place and drop them when they don't.
Most teams treat the backlog as the truth, then spend their energy forcing reality to match it. We do the reverse. What's actually happening is the truth, and the backlog and roadmap follow it. When the ground shifts, the tickets move.
- 08:03
- 08:21
- 08:36
- 08:52
- 09:10
- 09:41
- 09:43
- 10:25
- 11:02
- commit08:47
- commit09:12
- pull request09:38
- deploy09:57
- commit10:14
- commit10:31
- alert10:40
- pull request10:58
- deploy11:06
You can watch us work
We join your Slack and open a channel named after you: #yourcompany-log. It carries our running commentary on the work: what we're building, what we've just learned, what changed this morning, and the demo from ten minutes ago.
It's noisy by design. Mute it if you like. When you want to know where things really stand, it's there.
A second shared channel carries the machinery: commits, pull requests, deploys and alerts. Between the two, anyone can follow the whole project without asking.
Most organisations don't have this. Between updates, they're flying blind. We'd rather fly on instruments.
Focus and blur
We see a startup as a picture with some parts in sharp focus and the rest deliberately blurred.
In focus is the thing your first customer will pay for. We commit to it fully and build it properly.
Blurred is everything that might matter later, or might never matter. We defer it, buy it off the shelf, or build the simplest version that lets us move on.
When a blurred area comes into focus, because a customer needs it or the product demands it, we commit to it properly then. That isn't cutting corners. It's not spending your runway on things you may never need.
Checkpoints you can show people
We plan in checkpoints: moments where your product can do something new and meaningful that you can put in front of a customer, an investor or a co-founder. We agree them with you, work towards them relentlessly, ship eagerly, and adjust to what we learn.
Off-the-shelf first
If a managed service or an open-source tool does 80% of what's needed, we use it. We'd rather test an idea with something we wired together in a day than spend a month building it properly and find out nobody wants it.
When we do write bespoke code, we know exactly what it needs to do, because the off-the-shelf version taught us.
Hire one, get all of us
Each engineer you engage makes your company their major focus: your work comes first for them, and they carry the context. Around them, more of us follow your log channel, know your codebase and pitch in when a problem needs another head. Our Lead engineers stay close to every engagement.
You engage one engineer and get the whole team's experience, at no extra cost.
Everything in your name
We never host your infrastructure. Your cloud accounts, Google Workspace, GitHub and password manager are set up in your name, and the credentials are yours from day one.
If you ever stop working with us, there's nothing to hand back, because you already have it all.
Compliance when it counts
ISO 27001 and SOC 2 are worth doing exactly when a customer, a regulator or a market demands them, and not before. Early on, an audit regime only slows you down.
When it comes into focus, say because a big customer won't sign without it, we take it on properly and run it well. We're as comfortable passing an external audit as we are hacking together a prototype.
We're trying to make ourselves redundant
Our aim is for you to outgrow us. When you're ready to hire your own team, we help you find them. We hand over a tested, documented codebase with every decision written down, and we step back.
We call that graduation. It's the best measure of whether we did our job.
Why this is also why AI works for us
None of this was designed for AI. But a team that writes everything down, works in the open and treats reality as the source of truth turns out to produce exactly the context AI agents need.
It's why over 90% of our code is now written through AI sessions, and why it's still good code. As Will puts it: AI has made producing code cheap, but it hasn't made context cheap.
How we work AI-nativeRead Will's essay: Reality Is the Source of Truth
Want to see what working with us is actually like?
Tell us what you're building. If it's a fit, you'll be reading our log channel within a week.