Dremio MCP Server

# Framework Integrations (Experimental) The Dremio MCP server provides integrations with popular AI agent frameworks, allowing you to use the same MCP tools with different agentic frameworks. Currently supported frameworks are [LangChain](https://python.langchain.com/) and [BeeAI](https://github.com/i-am-bee/beeai-framework). BeeAI framework already supports MCP. ## Overview The frameworks integration enables: - Reuse of MCP tools across different agent frameworks - Consistent tool behavior regardless of the framework - Framework-specific optimizations and features ## LangChain Integration ### Features - Automatic tool discovery and registration - Integration with LangChain's ReAct agent - Support for multiple LLM providers (OpenAI, Ollama) - Structured tool responses ### Usage Example ```shell # Runs the LangChain server with all `tools` preconfigured in a simple # commandline interactive loop $ uv run python -m dremioai.servers.frameworks.langchain.server ``` Programmatic usage can be looked at in the [src/dremioai/servers/frameworks/langchain/server.py] ### Configuration Configure LangChain settings in your config file as mentioned in [settings](settings.md#langchain-settings-experimental) ## BeeAI Integration ### Features - MCP server integration with BeeAI agents - Sliding memory window support - Multiple LLM provider support - Environment variable management ### Usage Example ```shell # Runs the BeeAI server with the MCP server on the commandline in an interactive loop $ uv run python -m dremioai.servers.frameworks.beeai.server ``` Programmatic usage can be looked at in the [src/dremioai/servers/frameworks/beeai/server.py] ### Configuration Configure BeeAI settings in your config file as mentioned in [settings](settings.md#beeai-settings-experimental) ## See Also - [Tools Documentation](tools.md) - [Settings Configuration](settings.md) - [Architecture Overview](architecture.md)