Measured, Not Marketed

Effectiveness on a Real Production Codebase

A CRM platform (TypeScript/React/Node) before and after a Phase 4 rebuild. Every number traces to a neuralmind benchmark . command.

Token Reduction

48.8×

1,033 vs 50,000+ tokens/query

Wake-up Tokens

455

Per query, measured

Personal Edges

+275%

36 → 135 co-activations

Communities

810

Architectural boundaries

Before / After

Phase 3 vs Phase 4

Full rebuild on a real production CRM codebase

MetricBeforeAfterChange
Total nodes9,1909,293+103
Communities810new
Personal edges36135+275%
Shared edges10,07910,183+104
Shared edge weight2,774.732,924.66+5.4%
Wake-up tokens455
Avg query tokens1,033
Avg token reduction48.8×
Build Performance

5.4 min full rebuild, then ~30s increments

Full Rebuild

326s

All nodes, edges, embeddings

Incremental (after)

~30s

Changed files + dependents only

Honesty Gate

What we cannot claim yet

Self-improving

No production fitness data. Architecture is complete but gains are unmeasured.

Team memory that learns

No co-view signal yet. Structural seeds exist, but synaptic learning needs weeks of sessions.

Zero code egress

Overclaim. The agent layer still talks to its model. What NeuralMind itself does: transmits no repository content, and sends no telemetry.

See it in 30 seconds

Clone the repo, run the demo, get numbers on YOUR codebase. Then email the output to [email protected] with your team size for a free full spend model.

Run the demo