mw75-streamer 1.0.1

Stream EEG data from MW75 Neuro headphones using BLE and RFCOMM

MW75 Neuro Streamer

Stream 12-channel EEG data from MW75 Neuro headphones with WebSocket, CSV, and LSL output support.

About uv: This project uses uv for fast, reliable Python package management. Benefits include faster installs, better dependency resolution, and reproducible environments. All commands can be run with regular Python too (see Alternative: Using Python Directly), but we use uv throughout this documentation for consistency.

Features

• Real-time streaming: 500Hz, 12-channel EEG with µV precision
• Multiple outputs: WebSocket JSON, CSV files, Lab Streaming Layer (LSL)
• Built-in testing: WebSocket servers with browser visualization
• Robust protocol: Checksum validation and error detection

Installation

Option 1: Install from PyPI (recommended)

uv pip install mw75-streamer

For additional features (WebSocket, LSL support):

uv pip install "mw75-streamer[all]"

Option 2: Install from source

# Clone this repository
git clone https://github.com/arctop/mw75-streamer.git
cd mw75_streamer

# Install uv if needed (see installation guide: https://docs.astral.sh/uv/getting-started/installation)
brew install uv

# Create environment and install package
uv venv && uv pip install -e ".[all]"

Usage

# Basic streaming
uv run -m mw75_streamer --browser
uv run -m mw75_streamer --csv eeg.csv
uv run -m mw75_streamer --ws ws://localhost:8080
uv run -m mw75_streamer --lsl MW75_EEG

# Combined outputs
uv run -m mw75_streamer --csv eeg.csv --ws ws://localhost:8080

Developer Examples

For advanced integration into your own applications, see the examples/ folder:

• simple_callback.py - Quick start example for basic callback usage
• callback_integration.py - Comprehensive example showing real-time EEG processing using custom callbacks
• threaded_processing.py - Threading patterns for heavy processing (recommended for ML/filtering)
• Custom Callbacks: Process EEG packets, raw data, and events directly in your code
• Performance Guidance: Keep callbacks fast (< 1ms) or use threading for heavy work
• Integration Patterns: Combine callbacks with existing outputs (CSV, WebSocket, LSL)

# Quick callback example
from mw75_streamer import MW75Streamer, EEGPacket

def process_eeg(packet: EEGPacket):
    # packet.channels = 12 EEG channels in µV
    print(f"Ch1: {packet.channels[0]:.1f} µV")

streamer = MW75Streamer(eeg_callback=process_eeg)
await streamer.start_streaming()

See examples/README.md for complete documentation.

Testing

# 1. Start test server
uv run -m mw75_streamer.testing --advanced
# Optional: Press 'b' + Enter in server terminal to open browser visualization

# 2. Start EEG streaming
uv run -m mw75_streamer --ws ws://localhost:8080

How It Works

1. BLE Activation: Discovers MW75 via Bluetooth LE and sends activation commands
2. RFCOMM Streaming: Connects to channel 25 and receives 63-byte packets
3. Data Processing: Converts raw ADC to µV, validates checksums, outputs to CSV/WebSocket/LSL

Data Formats

CSV: Timestamp,EventId,Counter,Ref,DRL,Ch1RawEEG,...,Ch12RawEEG,FeatureStatus

WebSocket JSON: Real-time streaming with timestamp, counter, ref/drl, and 12 channel values in µV

Requirements

• Hardware: MW75 Neuro headphones (paired via Bluetooth)
• OS: macOS (fully supported), Linux (planned - contributions welcome)
• Python: 3.9+

macOS Setup for LSL

# Install LSL library (for LSL support)
brew install labstreaminglayer/tap/lsl
export DYLD_LIBRARY_PATH="/opt/homebrew/lib:$DYLD_LIBRARY_PATH"

# Pair MW75 headphones in System Preferences > Bluetooth

Performance Optimization

For improved real-time performance and reduced packet drops, run with elevated priority:

# Run with high priority (requires sudo for optimal performance)
sudo uv run -m mw75_streamer --csv eeg.csv

# The streamer automatically sets:
# - Process priority (niceness -10)
# - Thread real-time scheduling policy
# - Optimized RFCOMM event loop timing (1ms intervals)

Note: Running without sudo will still work but may have higher packet drop rates under system load.

Troubleshooting

• MW75 not found: Ensure headphones are powered on and paired
• Connection failed: Re-pair device in Bluetooth settings
• Dropped packets: Reduce Bluetooth interference, move away from WiFi routers and other 2.4GHz devices

Alternative: Using Python Directly

All uv commands can be replaced with regular Python. Simply activate your virtual environment first:

# Example: Replace 'uv run -m mw75_streamer' with 'python -m mw75_streamer'
source .venv/bin/activate
python -m mw75_streamer --csv eeg.csv --ws ws://localhost:8080
python -m mw75_streamer.testing --advanced

# Or replace 'uv pip install' with 'pip install'  
pip install mw75-streamer

Development

See CONTRIBUTING.md for development setup and contribution guidelines.

License

MIT License - see LICENSE for details.

About

MW75 EEG Streamer was developed by Arctop, a neurotechnology company focused on making brain-computer interfaces accessible and practical.

Acknowledgments

AI Assistance

• Claude Code (by Anthropic) - AI coding assistant used for development support and code optimization.

Open Source Dependencies

This project builds upon excellent open source libraries:

• bleak - Cross-platform Bluetooth Low Energy library for Python
• PyObjC - Python bridge to Objective-C for macOS integration
• websocket-client - WebSocket client library for real-time streaming
• websockets - WebSocket server implementation for testing tools
• pylsl - Python bindings for Lab Streaming Layer
• black - Python code formatter for consistent style
• mypy - Static type checker for Python
• flake8 - Python linting tool for code quality

Hardware & Community

• Master & Dynamic for creating the MW75 Neuro headphones and making EEG accessible
• The Python community for excellent Bluetooth libraries and frameworks

---

For detailed technical information about the MW75 protocol, see the inline documentation in the source code.