Selected work, in detail
Case studies
Every number on these pages was counted, not estimated. Where a case is our own build rather than client work, it says so at the top.
How we build — web and mobile in one repository
One repository, three releases
A web console, an iOS app, an Android app and a database underneath. Most teams run that as four codebases and four release days. This is the setup we build instead — one repository, two branches, and a push that releases all of it at once. It is running in production right now, and it is the same shape we would set up for you.
Read the case →3
systems released by a single push
2
complete environments, preview and production
Reference build — our own collection, published
Seventy-six automations, yours to take
Most automation pitches show you a five-node demo. This is the opposite: the workflows we actually built and ran, exported exactly as they are and put in a public repository. Import the JSON, plug in your own credentials, keep whatever is useful.
Read the case →76
workflows in the repository
57
of them call a language model
Reference build — our own setup, one user
A fix from a moving train
Development normally waits for the laptop. You see the problem on a phone, and nothing can happen until you are back at a desk. So we moved the agent off the laptop entirely: it lives on a small server, it is awake all the time, and you reach it by sending it a message like you would a colleague.
Read the case →1
person the bot will answer
0
laptops open when the fix went out
Reference build — our own system
The knowledge base that files itself
Everyone wants the second brain everyone on tech Twitter keeps describing. We built ours and have run it every day since April. An agent reads each source, writes the page, extracts the people and the ideas, and connects them — so the context is there before you ask for it.
Read the case →749
pages the agent maintains
369,654
words of context already written up
Reference build — demo on synthetic documents
A deal room where the AI never sees the internet
Law firms and advisories keep asking the same question: can we get what ChatGPT does without handing client documents to someone else's servers? We built the answer as something you can watch — a private deal room running on an open-source model on our own GPU.
Read the case →0
tokens sent to a third-party API
5
facts pulled from the documents, each cited
