We turn difficult information into software people can use. From clinical records and connected knowledge to market analysis and independent research.
Different problems. One instinct: bring structure to complexity and make the next step clearer.
Search your own documents with the network unplugged.
An open-source engine that indexes a folder of files on ordinary CPUs and finds the right page by keyword and by meaning. Nothing leaves the machine unless you choose: a language model is optional, and Claude can use it as a tool.
View on GitHub Read the launch post
Turn scattered research into connected knowledge.
An AI-assisted workspace that compiles sources into a structured wiki, reveals connections in a knowledge graph, and lets people ask questions across their own material.
View demo
Make medical records easier to understand.
A medical information platform exploring how AI can organise dense records into clearer, referenced summaries for patients and professionals.
View demo
Test a trading idea before putting it to work.
A browser-based environment for exploring BTC strategies with historical market data, adjustable parameters, charts and side-by-side comparison.
Explore projectIndependent research, writing and experiments.
An independent blog researching software and tools for UK sole traders.
Hands-on reviews and comparisons for choosing AI voice tools.
A playful, interactive take on Conway's cellular automaton.
Our work moves between disciplines, but the method stays grounded in the people and decisions each system needs to support.
Start with the messy workflow, the scattered information and the decision someone needs to make.
Give information a useful structure, then design an interface that makes it easier to act on.
Keep improving the product around clear sources, thoughtful constraints and real-world use.
We welcome conversations about thoughtful software, product collaborations and difficult information problems.
Let's talk