PyneSys LLC developed the PyneCore library, which allows users to program trading strategies using Python instead of the Pine Script language used by TradingView. While Python-based development is supported, users must pay for a cloud-based compiler if they wish to use Pine Script.
To build the library, the developer requires a dataset of at least 2,000 popular trading strategies from TradingView. The process involves using an autonomous AI agent to reverse-engineer these strategies, identify discrepancies in calculations, and document them in a specific file.
The AI agent operates under strict constraints, including a limit of 120 compilations per hour and 2,100 per week via the PyneComp tool. The agent is tasked with documenting differences between PyneCore and Pine Script, which are then recorded in a technical log for potential submission as bug reports.
The workflow requires the agent to verify scripts, run tests, and maintain compatibility standards. The system is designed to automate the creation of a strategy constructor by analyzing existing code snippets, as individual strategies often lack sufficient performance on their own.
