txtai is an all-in-one open-source AI framework that combines semantic search, LLM orchestration and language-model workflows in a single Python library. Instead of stitching together a vector database, a search layer and an agent framework, you get all three in one coherent API, and the project has earned about 13,000 stars on GitHub.
At txtai’s core is an embeddings database that indexes text, documents, audio and images into vector, graph and relational stores at once, enabling hybrid queries that mix semantic similarity with structure and metadata. On top of that index sit pipelines for the common language tasks โ question answering, summarization, transcription, translation, text extraction and entity recognition โ and an LLM layer that supports agents, retrieval-augmented generation and multi-model workflows with providers ranging from local models to hosted APIs. Workflows can be declared in YAML as well as code, and everything can be exposed through a REST API, which makes txtai as comfortable inside a quick script as in a production service. Development has been continuous since 2020, and the framework remains a popular choice when a project needs serious search plus LLM features without adopting three separate platforms.
txtai is free and open source under the Apache-2.0 license. You run it on your own infrastructure; the only costs are your compute and any hosted model APIs you choose to call.
txtai is for developers who need semantic search plus LLM workflows in one framework โ for example document intelligence, support search or research tools โ and prefer a single well-integrated library over assembling a vector database, a search engine and an agent toolkit separately.