Shelly OS Β· data model and event bus

Starting Shelly OS







  Shelly OS Β· v Β· synthetic data
  
One brain. Every department talking.

  
Every department now reads one data model, and an event bus carries what happens to the agents who should act on it. A store stock-out raises the order and books the Gateway delivery. Small, routine actions run on their own; anything with money, people or customers waits for a person.









  
Event bus Β· today
What happened, which agents acted, and who has to sign off

  

  






  
Departments talking
One business's event, another business's action

  






  
One data model
The same tables and keys for every business Β· SQLite Β· real data stays on your PC

  






  
How it works

  
    
    
  1. Sense. Detectors read the data model and publish events: store.stock.below_cover, gateway.service.below_target, totara.expiry.at_risk.
    2. 
    
  2. React. Every agent subscribed to that kind of event proposes an action, in its own department, even in another business.
    3. 
    
  3. Check. Each action meets the agent's autonomy and the Brain's approval chain: department head, then a Finance funds check, then Pavi.
    4. 
    
  4. Log. Events and actions go back into the model, so every page and every agent sees the same story.
    5. 
  

  
Run it yourself: python -m shelly os run (today's events) Β· python -m shelly os tables Β· python scripts/release.py 2.0.1 ships a release in one step.