MCP Playground
Facilitates chat interactions with LLMs and seamlessly integrates external MCP tools using a Streamlit-based interface.
Acerca de
MCP Playground provides a Streamlit-based environment for interacting with Large Language Models (LLMs) while seamlessly integrating external tools via the Multi-Server Command Protocol (MCP). By leveraging LangChain and LangGraph, the playground offers provider-agnostic support for models from OpenAI, Amazon Bedrock, Anthropic, and Google. The Docker-first approach ensures easy setup and scalability, while the extensible architecture allows for the addition of new MCP servers and providers without modifying the UI code.
Características Principales
- Provider-agnostic design supports various LLMs via LangChain.
- Supports multiple MCP servers for diverse tool integrations.
- 3 GitHub stars
- Docker-based deployment for easy setup and scalability.
- Offers a rich Streamlit chat UI with history and live tool execution output.
- Uses LangGraph for dynamic tool selection and reasoning.
Casos de Uso
- Integrating custom tools into LLM workflows.
- Experimenting with different LLM providers and their capabilities.
- Developing and testing agentic AI applications.