Semango FAQs

Question 1

What is Semango and what does it do?

Accepted Answer

Semango is a hybrid search engine designed to answer natural language queries across diverse content. It intelligently combines traditional lexical (BM25) and advanced semantic (vector) search methods for highly relevant results.

Question 2

Does Semango support local embedding models for privacy or offline use?

Accepted Answer

Yes, Semango is flexible. It supports the OpenAI API for embeddings but also allows you to use local ONNX models (e.g., all-MiniLM-L6-v2), providing options for privacy-sensitive applications or environments without internet access.

Question 3

What types of content can Semango index?

Accepted Answer

Semango supports multi-format ingestion, allowing you to index a wide range of content including markdown, code files, PDFs, images, and various tabular data formats like CSV, JSON, Parquet, and SQLite.

Question 4

How can users interact with Semango?

Accepted Answer

Semango offers multiple interaction methods: an embedded React-based web UI with dark mode for easy browsing, and a token-authenticated REST API for programmatic access and integration into other applications. It's also distributed as a single binary for simple deployment.

Question 5

What is hybrid search, and why is it important in Semango?

Accepted Answer

Hybrid search in Semango combines the strengths of lexical (keyword-matching via BM25) and semantic (meaning-based via vector similarity) search. This fusion delivers more accurate and comprehensive results, especially for complex natural language queries where a simple keyword match might miss context.

Semango

Semango

About

Key Features

Use Cases

About

Key Features

Use Cases