NMX-NeuroMesh

NMX-NeuroMesh is an experimental Python prototype that combines RF-derived input, JAX-accelerated nonce scoring, hash generation, and lightweight UDP broadcast signaling in a blockchain-style node loop.

The repository currently contains one node implementation. It is a research sandbox, not a finished cryptocurrency client or production blockchain.

Repository contents

• neuromesh_node_v13.py — the Phase VII prototype, with generated RF-like input as a fallback and optional LimeSDR RX/TX through SoapySDR
• LICENSE — the MIT License
• README.md — project overview and usage notes

How the prototype works

At a high level, neuromesh_node_v13.py:

1. opens UDP port 5005 and listens for JSON node announcements;
2. initializes a LimeSDR through SoapySDR when compatible hardware and drivers are available, otherwise uses generated random samples;
3. derives a simple entropy value from the input;
4. scores batches of nonce candidates with JAX;
5. hashes the selected nonce and score;
6. broadcasts the resulting node/hash announcement over UDP; and
7. transmits an RF burst when the SDR path is active.

The prototype does not implement durable consensus, authenticated peers, chain validation, persistent storage, wallets, or production-grade cryptography.

Requirements

• Linux-like environment
• Python 3
• numpy
• jax
• jaxlib

Optional RF support requires:

• SoapySDR
• LimeSDR hardware and the matching driver stack

Safety before running

neuromesh_node_v13.py is not receive-only. If it discovers a compatible LimeSDR, it configures RX and TX at 915 MHz, sets TX gain to 60, activates both streams, and sends a burst after finding each block. It also binds UDP port 5005 and broadcasts node announcements on the local network.

Before executing it:

• comply with the radio regulations and band plan for your location;
• use a shielded, properly attenuated loopback setup for bench testing;
• do not attach a radiating antenna unless the frequency, power, authorization, and operating conditions are lawful and intentional; and
• disconnect or disable compatible SDR hardware if you only want to inspect the generated-data fallback.

The current script has no command-line option to disable TX.

Usage

Create an isolated environment and install the core Python dependencies:

python3 -m venv .venv
source .venv/bin/activate
python3 -m pip install --upgrade pip
python3 -m pip install numpy jax jaxlib

After reviewing the safety notes and the source:

python3 neuromesh_node_v13.py

Stop the loop with Ctrl-C.

If SoapySDR or LimeSDR initialization fails, the script reports [RF] FALLBACK and uses generated input. This fallback still performs UDP network activity.

Development status

NMX-NeuroMesh demonstrates an experimental AI/RF/hash workflow. Natural next steps include separating RF, networking, and scoring into modules; adding a TX-disable option; defining validation and persistence rules; documenting a reproducible hardware configuration; and adding automated tests.

License

Released under the MIT License.