FluctlightDB

Type a cue. Watch it spread.

This page talks to one ordinary Python process with the engine embedded inside it, the same call you would make locally after pip install "fluctlightdb[native]". Your brain is created when you arrive and thrown away when you leave.

Waking the engine.

Load a memory set

Or paste your own

A chat export, notes, anything. It is turned into memories in your own brain and discarded with it. Nothing is stored, and nothing is sent anywhere but this page's engine.

Load a memory set to begin.

Node size is salience. Depth is age. Edges are shared neuron populations. A ring means the memory carries a verified source.

What is real on this page

The memories, the graph, the ranking and the consolidation all come out of the engine. The edges are recovered from the neuron populations each memory encoded into, which is how the engine relates them.

The travelling part of the animation is not. activate() computes its spreading frontier and discards it before returning, on every surface the engine exposes, so no call can report it. What you watch move is a reconstruction across real edges, between the engine's real seeds and its real final ranking. We would rather say that than let a picture imply otherwise.

Source · Benchmarks · Reproducibility