Semango Hybrid Search Engine

Example Use Cases

• 1

  Perform natural language search over codebases and documentation.

• 2

  Integrate hybrid lexical and semantic search capabilities via MCP protocol.

• 3

  GPU-accelerated local embeddings or OpenAI-compatible embedding APIs.

Description

Semango is a hybrid search engine combining BM25 lexical search with vector similarity to index and search codebases and documents using natural language queries. It supports multiple file formats, incremental indexing, and embedding providers including OpenAI-compatible APIs and local ONNX models with GPU acceleration. As an MCP server, it allows large language models to access indexed content as a search tool via stdio or SSE transports, enhancing LLM tool-use capabilities.