$GWH × Bittensor Integration Roadmap
Detailed Steps
Install Prerequisites
Set up your development environment
Before connecting to Bittensor you need Python 3.8+, the Bittensor SDK, and a working knowledge of command-line tools. This step installs everything you need.
aInstall Python 3.8+
Bittensor requires Python 3.8 or higher. Check your version and install if needed.
python3 --version
# If < 3.8, install via your package manager:
# macOS: brew install [email protected]
# Ubuntu: sudo apt install python3.11 python3.11-venvbCreate a virtual environment
Isolate your GWH subnet dependencies from your system Python.
python3 -m venv gwh-subnet-env
source gwh-subnet-env/bin/activate # Linux/macOS
# or gwh-subnet-env\Scripts\activate # WindowscInstall the Bittensor SDK
The SDK provides btcli (wallet management) and the Python library for building neurons.
pip install bittensor>=8.0.0dInstall GWH subnet package
Clone the GWH subnet repo and install it in editable mode so you can customize agents.
git clone https://github.com/genesis-world-health/bittensor-subnet.git
cd bittensor-subnet
pip install -e .💡 Pro Tips
- →Use a dedicated machine or cloud VM (Ubuntu 22.04 recommended) for running neurons 24/7.
- →Ensure you have at least 8 GB RAM and a stable internet connection.
- →Consider using tmux or screen so your processes survive SSH disconnects.
✅ Outcome
You now have Python, Bittensor, and the GWH subnet code installed and ready.
Create a Bittensor Wallet
Generate your coldkey and hotkey
Register the GWH Subnet
Claim your NetUID on Bittensor
Configure Environment
Set API keys, wallet paths, and subnet parameters
Deploy a GWH Miner
Start earning $GWH by serving health recommendations
Deploy a GWH Validator
Score miners and secure the health AI network
Mainnet Launch & Token Economics
Go live with real $GWH distribution
🏗️ Architecture Reference
The GWH subnet maps the platform's existing 75+ AI health agent architecture to Bittensor's miner/validator paradigm.
| GWH Platform | Bittensor Subnet |
|---|---|
| UserHealthContext | HealthQuerySynapse (query fields) |
| AgentRecommendationOutput | HealthQuerySynapse (response fields) |
| FiveGatesValidation | five_gates_score() in reward.py |
| BaseAgent.processWithGateValidation() | Validator assess_response_gates() |
| AgentOrchestrator | Validator forward() + weight-setting |
| 14 Basic Health Agents | AGENT_MAP in miner.py |
| Subscription Tiers | Miner specialization categories |
⚖️ Five Gates Scoring Weights
Every miner response is scored across five ethical dimensions. The Do No Harm gate carries the highest weight — responses that fail it are hard-capped at 0.1 reward.