Agentic project hub
Erich OS
AI tools can do useful work, but a conversation is a poor long-term home for everything a project needs. I built Erich OS as a hub where projects, tasks, and AI-assisted execution can stay connected.
- My role
- Creator & system designer
- The work
- Agentic project hub
The problem
Work spread across repositories, tools, and conversations is easy to lose track of. A project needs a durable account of its purpose, current decisions, next actions, and who is responsible. An agent also needs a bounded way to contribute without becoming the owner of every decision.
What I built
Erich OS brings project records, tasks, links, people, and decisions into one working system. AI tools use their own grants to read context and make supported updates. A separate work flow lets me approve execution, with Rider providing a bridge to activity on my Mac. The interface makes it possible to move from a portfolio of projects into the details of a particular piece of work.
The result
The result is my working agentic hub: shared project context, accountable task updates, and a place to review and authorize work. It is a private operating tool. This case study presents the system and its purpose without exposing the project records inside it. Its next test is always practical: does it make the work easier to understand and move forward?