Reliable infrastructure for the agentic era.

Noise-canceling context and long-term memory for AI agents. Cuts a real command mix by 58.9%, leaves JSON and YAML untouched, and passes failing commands through verbatim so no error ever gets compressed away.

Point Claude Code at any LLM. A Go proxy that routes your agent traffic to whichever model you actually want to pay for.

Tiny macOS menu-bar monitor that names the app making your Mac heavy and quits it in one click. Apple Silicon & Intel, no daemon.

An AI-guided workflow for founders. Write down the assumption, ship the smallest thing that tests it, then watch what users actually do instead of what they said they would.

Reads the diff and writes the pull request description. Works with any OpenAI-compatible model, installs from the GitHub Actions marketplace in one step.

One set of colour, type and spacing tokens behind Omni, Bubo and this site. Three themes fill the same fifteen roles, and a test computes every contrast ratio and fails the build when one drops below WCAG.