# microsoft/PIKE-RAG

PIKE-RAG: sPecIalized KnowledgE and Rationale Augmented Generation

Repository: https://github.com/microsoft/PIKE-RAG
Canonical: https://ross.abutalabs.com/products/pike-rag
Homepage: https://arxiv.org/abs/2501.11551
Language: Python
License: MIT
License Family: permissive
Topics: domain-specific, industrial-ai, knowledge-extraction, rag
Last push: 2025-09-10T09:10:28+00:00

## Health v2 (maintenance only)
Score: 31/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 41, release rhythm 8, longevity 50
- inputs: {"age_days": 706, "days_push": 357, "days_rel": 586, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2480, forks 237 (observed 2026-08-28T04:06:55.266544+00:00)

## What it is
PIKE-RAG is a Microsoft framework for building Retrieval-Augmented Generation systems that extract, understand, and apply specialized domain-specific knowledge with coherent multi-step reasoning. It provides modular pipelines covering document parsing, knowledge extraction and storage, retrieval, knowledge-centric reasoning, and task decomposition.

## Use cases
- build a rag pipeline over industrial or domain-specific corpora
- extract specialized knowledge from professional documents for llm question answering
- answer multi-hop questions requiring reasoning over a document corpus
- parse pdfs and extract knowledge atoms for retrieval
- classify rag task complexity and phase rag system development
- search patient historical medical records with factual retrieval
- customize rag modules for different retrieval and reasoning capabilities

## When to choose
- you need rag beyond simple retrieval, with domain knowledge extraction and logical reasoning
- your corpus is specialized (industrial, medical, legal) and generic rag underperforms
- you want a modular, configurable rag pipeline in Python backed by research

## When to avoid
- you need a simple plug-and-play rag over general web content
- you want a hosted commercial product rather than a framework you assemble
- your use case is unrelated to llm-based retrieval and reasoning

## Facets
- artifact type: library
- maturity: active
- function: rag, nlp, llm-inference, machine-learning, parser
- domain: large-language-models, artificial-intelligence
- platform: python
- tags: knowledge-extraction, domain-specific, industrial-ai, knowledge-atomizing, task-decomposition, reasoning, retrieval-augmented-generation, natural-language-processing

## Member repositories
- microsoft/PIKE-RAG (main) score 31

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:55.266544+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:28:16.121616+00:00, confidence not recorded.
  - readme: https://github.com/microsoft/PIKE-RAG (fetched 2026-08-28T04:06:55.266544+00:00, sha c22f07430589)
  - homepage: https://arxiv.org/abs/2501.11551 (fetched 2026-08-29T10:10:31.139207+00:00, sha 62c85e1e4c07)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T10:10:31.148289+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T10:10:31.151671+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T10:10:31.153525+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T10:10:31.149983+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.