BlueChips is an integrated digital asset analysis platform that leverages STARK-verifiable technology to generate verifiable predictions for BTC and ETH market movements. The platform consists of three main components:
- xgb2cairo: A utility that converts trained XGBoost models to Cairo code for STARK-verifiable applications.
- BlueChipAI API: A REST API that provides BTC and ETH price predictions based on the Cairo-transpiled models.
- auto-swap-runner: A background service that periodically queries the prediction API and automatically executes asset operations using the AVNU SDK when positive signals are detected.
A tool that transpiles trained XGBoost models to executable Cairo code, specifically designed for BTC and ETH price prediction models.
Key Features:
- Converts XGBoost models to Cairo code with Fixed Point 32x32 representation
- Generates Scarb project structure for Cairo execution
- Includes validation tools to ensure model accuracy after transpilation
Technologies:
- Python 3.11.5
- XGBoost 2.1.3
- Cairo 2.11.4
- Scarb 2.11.4
A Python-based REST API that integrates external market data with Cairo-based models to provide verifiable, proof-generating predictions for BTC and ETH buy signals in real-time.
Key Features:
- Fetches price time series data for BTC and ETH from external sources
- Processes data through STARK-verifiable ML models implemented in Cairo
- Exposes prediction endpoints through a REST API
API Endpoints:
/api/v1/- Health check endpoint/api/v1/predictions/<asset>- Get predictions for a specific asset (BTC or ETH)/api/v1/predictions/all- Get predictions for both BTC and ETH/api/v1/predictions/<asset>/features- Get the lagged return features for an asset
Technologies:
- Python 3.12+
- Cairo 2.11.4
- Scarb 2.11.4
A lightweight, resource-efficient background service in TypeScript that queries the BlueChipAI prediction API and, upon receiving an affirmative signal, executes a swap action using the AVNU SDK on StarkNet.
Key Features:
- Runs as a background process, waking up only at configured intervals
- Configurable through environment variables
- Minimal resource usage with no busy-waiting
- Transparent logging of all operations and errors
Technologies:
- Node.js
- TypeScript
- AVNU SDK for StarkNet
Learn more about auto-swap-runner
- Python 3.11+ (for xgb2cairo and BlueChipAI)
- Node.js and pnpm (for auto-swap-runner)
- Cairo 2.11.4 and Scarb 2.11.4
curl --proto '=https' --tlsv1.2 -sSf https://docs.swmansion.com/scarb/install.sh | sh -s -- -v 2.11.4
-
Clone the repository:
git clone <repository-url> cd bluechips
-
Set up each component following their individual README instructions.
- Model Transpilation: Use xgb2cairo to convert trained XGBoost models to Cairo code.
- API Deployment: Deploy the BlueChipAI API which uses the transpiled Cairo models.
- Automated Operations: Configure and run the auto-swap-runner to periodically check for market signals and execute asset operations.
IMPORTANT: This software is for educational and research purposes only. The information and code provided in this repository is not intended to be and does not constitute financial advice, investment advice, trading advice, or any other type of advice. The predictions generated by this software should not be used as a basis for any investment decisions. Digital assets are volatile and speculative, and all trading and investment decisions are the responsibility of the individual.
The creators and contributors of this project do not assume any responsibility for financial losses or other damages resulting from the use of this software. By using this software, you acknowledge and agree that you are doing so at your own risk.
This project is licensed under the MIT License.