PlanningCopilot is a tool-augmented LLM system for the PDDL planning process, designed to run planning tools through natural language instructions, improving reliability without the need for domain-specific training.
Purpose: PlanningCopilot integrates multiple planning tools into a single chatbot, allowing users to run these tools with natural language instructions and complete the full PDDL planning process from planning solving and validation to execution.
Main capabilities:
- Built on the Model Context Protocol (MCP), making it easy for language models to interact with external tools and systems.
- Modular design, where each part can be independently replaced, upgraded, or extended without affecting the rest of the system.
- In the current implementation, Solve uses FastDownward for classical planning and Metric-FF for numeric planning; Verify uses VAL to validate plans; Execute relies on PDDL_Plus_Parser to simulate and track plan execution.
How to use:
- Make sure Python 3.10 is installed and activated (via a virtual environment or conda environment).
- Install the latest version of Ollama to run locally.
- Install all project dependencies.
- Update all paths and settings in the config.py file.
- Run the LLM chat.
- To change the LLM, edit the llm_with_tools.py file.
- To add new tools, modify the MCP server in solvers_server.py.
Applicable scenarios: Suitable for research or development scenarios that need to drive the PDDL planning toolchain through natural language, especially for users who want to integrate planning solving, validation, and execution tools in a modular way.
Limitations: The materials do not provide more details about fees, pricing, platform support scope, specific performance metrics, or usage steps; this information is not currently available.
