# Marker-Inc-Korea/AutoRAG

AutoRAG: Now your agent can find anything in your computer. It gets smarter if you are using it frequently.

Repository: https://github.com/Marker-Inc-Korea/AutoRAG
Canonical: https://ross.abutalabs.com/products/autorag
Homepage: https://marker-inc-korea.github.io/AutoRAG/
Language: TypeScript
License: NOASSERTION
License Family: other
Topics: analysis, automl, benchmarking, document-parser, embeddings, evaluation, llm, llm-evaluation, llm-ops, open-source, ops, optimization, pipeline, python, qa, rag, rag-evaluation, retrieval-augmented-generation
Last push: 2026-08-26T18:32:27+00:00

## Health v2 (maintenance only)
Score: 93/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 99, longevity 69
- inputs: {"age_days": 966, "days_push": 7, "days_rel": 9, "gap_med": 7.5, "n_releases_24m": 31}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5055, forks 432 (observed 2026-08-28T04:09:09.126472+00:00)

## What it is
AutoRAG 2.0 is a self-evolving librarian agent that searches local document collections (PDFs, wikis, notes) and curates results into structured, numbered knowledge units instead of raw search hits. The repository also maintains the legacy Python AutoRAG, an RAG AutoML tool for automatically optimizing RAG pipelines, in maintenance mode.

## Use cases
- search my pdfs and notes and give me an answer, not file paths
- build a librarian agent over my local knowledge base
- automatically find the optimal RAG pipeline for my data
- evaluate and benchmark RAG retrieval and generation quality
- create QA evaluation datasets for retrieval-augmented generation
- get smarter document search that learns from my feedback

## When to choose
- you want curated answers from local documents rather than raw search results
- you need an agent that improves retrieval strategy over time via memory
- you want to auto-optimize or benchmark RAG pipelines with the legacy tool

## When to avoid
- you need a simple keyword/grep search tool without LLM involvement
- you want a fully managed cloud RAG service rather than a self-hosted agent
- you require a permissive standard license (license is non-standard)

## Facets
- artifact type: application
- maturity: active
- function: rag, search-engine, agent-framework, llm-inference, nlp, pdf
- domain: artificial-intelligence, large-language-models, developer-tools, pdf
- platform: python, cli, cross-platform
- tags: rag-evaluation, automl, document-parsing, librarian-agent, self-evolving-memory, local-document-search, knowledge-curation, retrieval-augmented-generation, ai-agents, natural-language-processing, nodejs

## Member repositories
- Marker-Inc-Korea/AutoRAG (main) score 93

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:09.126472+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 2026-08-29T18:17:19.869677+00:00, confidence not recorded.
  - readme: https://github.com/Marker-Inc-Korea/AutoRAG (fetched 2026-08-28T04:09:09.126472+00:00, sha 80c027a045e8)
  - homepage: https://marker-inc-korea.github.io/AutoRAG/ (fetched 2026-08-29T08:58:01.845794+00:00, sha 755669831091)
- Data as of 2026-08-30T08:39:29.467469+00:00.