Live simulation · No CUDA required

Circuit lab

Loading toy…Browser CPU · no server simulation

Start with an artificial offline circuit, or explicitly download a real anatomical subgraph.

Preparing the simulation…

0 ms0 recorded commands

Grouped activity. Positions are schematic; lines show retained connections between groups.

Activity over time

Filtered firing rates · up to 250 samples · each sample follows 20 simulated milliseconds. Display speed depends on your device.

Application output

No output yet

This lamp maps the largest output intensity to brightness. It is an application effect, not biological movement.

0.000

Inspect a cell

Load a model to inspect its state.

Try a causal comparison

  1. Load the real circuit. Select looming_left, apply a 100 ms stimulus, then run. Observe the GF group and output.
  2. Reset, silence “All output neurons”, and repeat the same stimulus. Input populations can respond while new GF spikes are suppressed.
  3. Export the recording to reproduce the run in Python, or download a standalone HTML demo and edit its marked input/output section.

Recorded inputs are artificial current drives; this page does not simulate retinal images, natural escape behavior or learning. Read the source, assumptions and model card.

Replay a downloaded recording in Python · repository checkout
python examples/replay_experiment.py flybrain-experiment.json --download

Browser runtime source · Guide and Python parity checks