Project
OpenClaw Agents
A self-hosted agent fleet with its own app-building workflow.
What it is
My OpenClaw Agents are a self-hosted fleet of eight agents.
The main agent is a personal workhorse, researching topics daily and acting as admin for the overall system.
The fleet researches and semi-autonomously builds apps through its own SDLC workflow. Agents have different roles across implementation, research, security, documentation, UX, product, coordination, and assistant work.
How it works
The project is built around a working agent fleet rather than a single chatbot.
Agents coordinate through familiar messaging surfaces including Discord, Telegram, and WhatsApp, while the Mission Control dashboard gives a live overview of agent state, activity, health, sessions, jobs, and memory.
What it explores
The project focuses on agent systems that are useful, observable, and personally controlled rather than abstract demos. The important part is not only that agents can generate code; it is the orchestration layer around them: roles, workflow, memory, coordination, review, and the discipline of making semi-autonomous work observable.
Current status
Active and live.