txtai
neuml/txtai · Python · Apache-2.0
Embed, index, search and prompt — no separate vector DB to run.
txtai combines an embeddings index, semantic search, summarisation and LLM workflows into a single dependency, great for apps that want RAG without standing up a database.
Score
Setup at a glance
pip install txtai Deep dive
txtai is an all-in-one embeddings database: vector search with SQL, object storage, topic modeling and graph analysis in one lightweight Python library — no separate vector DB, orchestration engine and search stack to run.
It creates embeddings for text, documents, audio, images and video, then builds pipelines on top of them that run LLM prompts, question-answering, summarization, translation and transcription. Workflows join pipelines together and aggregate business logic, and agents intelligently connect embeddings, pipelines, workflows and external tools.
Everything is local-first and pip-installable, with a cloud option (txtai.cloud) when you want the same API managed.