txtai

txtai is a comprehensive AI framework designed for building semantic search applications, LLM orchestration, and complex language model workflows. At its core is an embeddings database that integrates vector indexes (sparse and dense), graph networks, and relational databases, providing a foundation for vector search and knowledge sources for Large Language Model (LLM) applications. Key capabilities include: - Vector Search: Supports SQL, object storage, topic modeling, graph analysis, and multimodal indexing for text, documents, audio, images, and video. - LLM Orchestration: Tools for building autonomous agents, Retrieval Augmented Generation (RAG) processes, and multi-model workflows. - Pipelines & Workflows: Pre-built pipelines for tasks such as question-answering, labeling, transcription, translation, and summarization, which can be joined into larger business logic workflows. - Agents: Intelligent agents built on the smolagents framework that connect embeddings and pipelines to solve complex problems. - API & Integration: Provides Web and Model Context Protocol (MCP) APIs with bindings for JavaScript, Java, Rust, and Go. It is built using Python 3.10+, Hugging Face Transformers, Sentence Transformers, and FastAPI.