ZepAI Memory Layer
Automatically converts FastAPI backend services into Model Context Protocol (MCP) tools and resources, acting as a flexible protocol adapter.
Acerca de
This server acts as an efficient bridge, transforming existing FastAPI backend endpoints into accessible Model Context Protocol (MCP) tools and resources. Designed for applications like a 'memory layer' backend, it significantly reduces development time and code complexity by auto-generating MCP interfaces from your FastAPI application. It enables seamless integration with MCP clients, offering functionalities for semantic search, diverse data ingestion (text, code, messages, JSON), and secure exposure of read-only administrative insights, all while maintaining a single source of truth for your API definitions.
Características Principales
- Semantic search capabilities for data retrieval with advanced reranking strategies.
- 0 GitHub stars
- Comprehensive data ingestion functionalities for various data types including text, messages, JSON, and code with LLM scoring.
- Combined access via MCP client and original FastAPI routes, including OpenAPI documentation.
- Automatic conversion of FastAPI endpoints into MCP tools and resources.
- Secure exposure of read-only admin endpoints as MCP resources, filtering out sensitive POST operations.
Casos de Uso
- Building knowledge graphs or conversation history services accessible via MCP.
- Exposing an existing FastAPI-based AI memory layer for context management to MCP clients.
- Integrating code change tracking and ingestion into an MCP-compliant agent workflow.
