A minimal interpretation engine that turns real-world signals into meaningful, replayable events.
Most systems log data.
This system tries to log what actually happened.
It takes raw signals (like electricity usage) and interprets them into structured events, then renders them as a human-readable timeline.
Not just numbers. Not just logs.
Memories.
Because raw data is not useful on its own.
- 120W means nothing
- 2 hours means nothing
But:
- "someone cooked"
- "AC has been running too long"
- "something abnormal happened"
That is meaning.
Signal → Event → Template → Narrative → Timeline
- Raw JSON = what the system sees
- Templates = how the system speaks
- Timeline = what actually happened (interpreted)
The core engine is intentionally small. Just a few lines of logic.
The complexity comes from the world, not the code.
- A minimal interpretation layer
- Event → narrative mapping
- Real-time story generation from signals
You are not looking at a dashboard.
You are looking at a system describing reality.
This is a small step toward a bigger idea:
A replayable world.
A system where:
- environments remember themselves
- events can be replayed as stories
- reality becomes a structured, queryable timeline
Not logs.
A memory layer for the physical world.
Experimental.
Probably overthinking things.
But also maybe the right direction.