Self-hosted infrastructure that lets a single operator move at the pace of a team — AI agents, task queues, and operational dashboards that actually run in production.
Two systems built for real workloads. Both run in production on a single VPS.
A web-first queue and dispatch layer for running Claude Code, Codex, and Gemini CLI agents from one workspace. Manages preflight evaluation, token budgets, spill routing, and execution history — so you never juggle terminals again.
Live instanceA Star Trek-inspired personal operations system that captures tasks from email, Teams, tickets, and meetings — triages them with AI, stores them in Obsidian, and nags you to completion via Telegram. Designed for operators managing too many inbound channels.
Live instanceA short list of things we actually believe, not a mission statement.
Your data, your compute, your costs. No SaaS dependency on the critical path.
Built for one human who needs leverage, not a team that needs consensus.
Agents are services. They get queues, logs, retries, and budgets — like everything else.
One VPS. Three services. No Redis required. Scale up when you prove the need.
A running system beats a perfect spec every time. Roadmaps exist to be revised.
TypeScript, Postgres, Docker. Choose proven tools and save creativity for the product.