Software should serve the human, not extract from them.
Fullstack Software Engineer with a passion for sovereign and open source systems.
AI-powered therapy notes that run entirely on-device. No cloud, no uploads, no training pipeline — progress notes in under a minute.
Measure and improve your coding-agent hygiene. 25 checks across context efficiency, cost, structure and habits — with letter-grade scoring and auto-fix.
Track token spend and model usage across your agent sessions in real time. Local-first, no telemetry.
Markdown-first agent orchestration. Reads your existing todo.md and activity-log.md, dispatches Claude Code sessions, watches them work, validates output.
Swap between local LLMs on demand from a simple desktop app. Keep models cached, pick the right one per task, stay off the cloud.
Don't hesitate to reach out with any questions or ideas. Your message could spark something wonderful - let's start a dialogue.