Built my own AI command centre in under 24 hours using Claude Code, Ollama & multi-agent workflows
Yesterday I had an idea I couldn’t stop thinking about:
What if a single dashboard could run multiple AI agents locally and in the cloud — each with different jobs, memory, tools and workflows?
So I sat down with Claude Code and started building.
Under 24 hours later, I had a working prototype running on my MacBook Air.
Current stack:
Claude Code as the primary orchestration layer
Ollama running Hermes locally
OpenClaw for multi-agent workflows
Node.js task runners
Background automation + shell execution
Local-first architecture
Current agents:
Claude Code → reasoning, orchestration, coding
Hermes → local/offline LLM tasks
OpenClaw → workflow chaining
Task Runner → scheduled jobs + shell tasks
The interesting part isn’t the UI.
It’s watching agents hand work between each other:
one summarises
another executes
another validates output
another schedules follow-up tasks
Basically a lightweight AI operations centre running on consumer hardware.
Still early.
Still rough.
But it already feels different from “just another chatbot wrapper.”
Curious where people think this space is going:
AI command centres?
local-first agent systems?
autonomous workflows?
personal AI infrastructure?
Would genuinely appreciate feedback from builders working on similar things.
Any advice or tips would greatly help me out!