Skip to content

AI and developer tools

Language models wired into products, and tools developers use every day.

We build with language models and for developers: assistants that work in a terminal, model integrations that are not tied to one provider, and small tools that remove friction from everyday work.

Where code runs on someone's behalf, it runs inside a sandbox with limits.

5 capabilities, each built in real projects

What we have built

Capabilities in this domain.

  • AI agents2 projectsTerminal assistants that read, change and test code.
  • LLM integration1 projectModels from any provider connected to a product.
  • Sandboxed execution1 projectRunning untrusted code safely, inside limits.
  • In-browser developer tools1 projectUtilities that run entirely in the browser, no sign-in.
  • GitHub API integration1 projectReading repositories, branches and files programmatically.

How we approach it

What you can expect from us here.

Provider-neutral
Model calls go through one interface, so changing provider is configuration, not a rewrite.
Sandboxed by default
Anything an agent runs has explicit limits on files, network and time.
Small tools stay small
Utilities that run entirely in the browser when that is all the job needs.

Next step

Tell us what you need built.

Describe the problem in your own words. We reply within one business day with questions, then a written scope and price. No commitment until you accept it.