microsoft/typeagent-py

Structured RAG: ingest, index, query

What it solves

TypeAgent is designed to implement structured RAG (Retrieval-Augmented Generation), providing a Pythonic translation of the TypeAgent KnowPro system to help users build more organized and structured retrieval systems.

How it works

It acts as a Python package that interfaces with third-party LLMs to process and index information for structured retrieval, translating the logic and functionality of the original TypeScript-based TypeAgent KnowPro project into Python.

Who it’s for

Developers who want to implement structured RAG patterns in Python and those interested in the experimental prototype of the TypeAgent ecosystem.

Highlights

  • Experimental prototype for structured RAG MVP.
  • Python translation of the TypeScript TypeAgent KnowPro project.
  • Easy installation via pip.

Related

  • Project
    microsoft/TypeAgent

    TypeAgent is a Microsoft open‑source sample that shows how to build a single personal AI assistant using large language models, structured prompting, and a novel “Structured RAG” memory system. It includes an Electron shell, a dispatcher that routes natural‑language requests to typed agents, a memory layer that stores conversation facts as logical entities, and a cache to reduce LLM calls. The repo provides many example agents (calendar, email, browser, etc.) and an SDK for adding custom agents. It is early‑stage sample code, tested with Azure OpenAI in English, and not intended for production without further validation.

  • Project
    weiwill88/Local_Pdf_Chat_RAG

    A modular Python reference implementation of a RAG pipeline that allows developers to learn and experiment with document parsing, hybrid retrieval, and LLM generation.

  • Project
    awslabs/graphrag-toolkit

    A collection of Python tools for building graph-enhanced Generative AI applications using hierarchical lexical graphs and structured knowledge graphs.

  • Project
    NirDiamant/Controllable-RAG-Agent

    A Python library that builds a deterministic‑graph‑driven RAG agent. It ingests PDFs, creates chapter summaries and vector stores, anonymises queries, plans multi‑step reasoning, retrieves and verifies information, and finally generates grounded answers. Includes a Streamlit visualiser, Ragas‑based evaluation, and works with OpenAI, Groq, or any LangChain‑compatible model.

  • Project
    GiovanniPasq/agentic-rag-for-dummies

    A modular Agentic RAG framework using LangGraph that implements hierarchical indexing, multi-agent parallel retrieval, and human-in-the-loop query clarification.