quantalogic/quantalogic

What it solves

QuantaLogic is a Python framework designed to simplify the creation of AI agents that can perform complex tasks, automate repetitive workflows, and engage in natural conversations. It bridges the gap between raw Large Language Models (LLMs) and practical applications by providing a structured way for agents to use tools, execute code, and follow predefined processes.

How it works

The framework operates through three primary modes of interaction:

• ReAct Mode: Uses a "Reason + Act" loop where the agent thinks step-by-step and uses a library of 40+ built-in tools (like web search and file management) to solve open-ended problems.
• CodeAct Mode: An advanced extension where the agent uses executable Python code as its primary action language, allowing it to generate and run code to solve mathematical or programming tasks iteratively.
• Flow Mode: A YAML-based workflow engine that allows users to define structured pipelines using nodes and transitions for repeatable automation.

It integrates with various LLMs (GPT-4o, Claude, DeepSeek) via LiteLLM and supports Docker-based isolation for safe tool execution.

Who it’s for

QuantaLogic is built for developers, business innovators, and AI enthusiasts who want to build agents capable of coding, data analysis, and business process automation without building the agentic logic from scratch.

Highlights

• Multi-modal Agent Logic: Supports ReAct, CodeAct, and Flow paradigms in one unified API.
• Extensive Toolset: Includes over 40 built-in tools and a plugin system for custom extensions.
• Code-First Execution: CodeAct enables agents to use Python as a primary tool for high-precision tasks.
• Structured Automation: Flow mode allows for the creation of complex, repeatable data pipelines via YAML.
• Safe Environment: Offers Docker-based isolation for executing agent-generated code.