Custom Bots
Bot templates and blueprints in the0
Custom Bots
Are can be though of as the blueprints for Bots. They are the blueprints that define the actual trading algorithm that will be executed by the Bots. Custom Bots are created using the The0 CLI and can be shared with other users through the Custom Bot Marketplace. They are defined in a structured way using YAML and JSON Schema, allowing for easy customization and deployment.
Understanding Custom Bots
Blueprint Architecture
Custom Bots serve as reusable programs that:
- Define the trading algorithm and logic
- Specify configuration parameters through JSON Schema
- Include all necessary dependencies and requirements
- Can be instantiated multiple times as different Bots
Key Components
- bot-config.yaml: The main configuration file defining the bot's metadata and structure
- Entry Points: The main execution functions for bot logic and backtesting
- Schemas: JSON Schema definitions for bot and backtest configuration
- Dependencies: Language-specific dependency files (requirements.txt, package.json)
Development Process
Custom Bots are:
- Created using either Python, JavaScript (more languages coming soon!)
- Developed locally with your preferred tools and frameworks
- Tested thoroughly before deployment
- Deployed to the platform using
the0 custom-bot deploy
Sharing and Monetization
Once created, Custom Bots can be:
- Kept private for personal use
- Published to the marketplace for free
- Monetized by setting a price in the marketplace
Benefits
- Reusability: Deploy the same Custom Bot multiple times with different configurations
- Version Control: Manage different versions of your trading strategies
- Collaboration: Share and sell your strategies to other traders
- Flexibility: Want to build ML bots with Python (with scikit-learn) or web3 powered bots with JavaScript? The0 supports them both!**
Note: Support for more languages coming soon! Watch this space 🎉
Note: There are some limitations on the libraries you can use your maximum code size is 1GB, so no pyTorch or TensorFlow yet! But you can use scikit-learn, pandas, and other lightweight libraries for classical machine learning. If you want to use heavy weight modules delegate that model to an infrence endpoint and make the bot query that
Related Terms
- Bots - The running instances of Custom Bots
- Marketplace - Where Custom Bots are shared and sold
- Bot Deployment - Creating Bot instances from Custom Bots