ArchestackBest of Open-Source AI: List on GitHub ↗3 stars (opens in a new tab)

#2222nd of 26 in RAG and knowledge

txtai

Python embeddings database with pipelines, workflows and agents for search and RAG

Documentation ↗ (opens in a new tab)Website ↗ (opens in a new tab)Repository on GitHub ↗ (opens in a new tab)
Stars
13k
License
Apache-2.0
Last commit
Oct 2026
Last release
Oct 2026
Language
Python

Missed the latest GitHub check; these facts may be older.

Overview

txtai is a Python library built around an embeddings database that combines sparse and dense vector indexes, graph networks and relational storage. On top of it sit LLM pipelines, workflows, RAG and agents (built on smolagents), exposed through a FastAPI web API and an MCP API. LLMs can come from Hugging Face, llama.cpp or LiteLLM.

Who it is for: Python developers building semantic search, RAG or agent apps

Strengths

  • Embeddings database combines vector indexes, graph analysis and SQL queries in one library
  • Pipelines for transcription, translation, summarization and labeling, not only LLM prompts
  • Web API from a YAML config via uvicorn; bindings for JavaScript, Java, Rust and Go
  • Over 70 example notebooks plus quickstart scripts for search, RAG, agents and workflows

Weaknesses

  • Code-first library with no built-in chat UI or visual workflow builder
  • No Docker or compose files detected in the repo; container docs are linked separately
  • Hardware requirements are not stated; the recommended LLM in the model guide is 31B
  • Agent layer is built on smolagents, so its behavior follows that upstream project

What it needs

  • GPU optional
  • Needs Python 3.11+, Hugging Face Transformers, Sentence Transformers, FastAPI
  • Models: Hugging Face models, llama.cpp, LiteLLM (OpenAI, Claude, AWS Bedrock)
  • port 8000

Also in RAG and knowledge

See all 26
Also in RAG and knowledge
RankProjectScore
1deeptutorSelf-hosted AI tutor with knowledge bases, guided learning, books and quizzes41.4k starsApache-2.080 out of 100
2Open NotebookSelf-hosted NotebookLM alternative with podcasts and 20+ model providers40.2k starsMIT76 out of 100
3LightRAGGraph-based RAG framework with a REST API server and web UI40.1k starsMIT76 out of 100
#4RAGFlowRAG engine with deep document parsing, template chunking and agentic retrieval Live demo ↗ (opens in a new tab)92.2k starsApache-2.074 out of 100
#5MaxKBEnterprise knowledge-base agent platform with RAG, workflows and MCP tools23k starsGPL-3.071 out of 100