Published on 2026-05-01 written by pepe and corrected by lluma
Metabolon — Entering the Thicket
In the previous texts, I introduced the idea of Thicket Web Mind (TWM) — a way of building systems not as monoliths or services, but as living constellations of small processes.
That might sound abstract. So let’s make it concrete.
Metabolon
On the surface, it is a simple system for tracking time, managing tasks, observing work. Let's call it a life organizer.
But it is also a playground where the runtime model is tested in real conditions.
It is not finished. No, it is not polished. But it is alive. And importantly — you can run it.
Entering the system
Getting Metabolon running is intentionally simple. The latest Janet and spork installed system-wide are required.
> git clone https://git.sr.ht/~pepe/metabolon
> cd metabolon
> janet bin/init.janet
> janet demiurge.janet
That’s it. You will set password, create configuration, and summon the Demiurge.
Control, not automation
Once the system is running, you can interact with it through the control script:
> janet bin/dm.janet state
> janet bin/dm.janet run-peers
> janet bin/dm.janet stop-peers
This is not an accident. The system does not start everything automatically. You are expected to inspect the state, decide when to run and control the lifecycle. This is a deliberate break from typical application design.
The operator is part of the system.What is actually running?
When you start Metabolon, you are not launching “an app”. You are starting a thicket where the Demiurge builds and supervises processes. The Tree holds the canonical state and Symbionts provide behavior and projections
All communication happens locally. The outside world only sees projections through HTTP/SSE membrane.
Why Janet
This model would be much harder in most languages. Janet makes it natural as code and data share the same structure, processes are small and cheap, so the runtime stays transparent. I stayed with it.
Current state
Metabolon is not a product. It is a system under construction. Yes things are rough with still evolving decisions. But the core is already visible.
What comes next
This text is not the explanation. It is the entry point. Next will be the user manual of the application. UI/UX and more. From there, I will start zooming in:
- the configuration and the Demiurge
- mycelium and membrane
- the Tree as canonical state
- symbiont lifecycles
- event flow
- information field navigation
- ... and beyond
Each part deserves its own space.