AutoRAG logo

AutoRAG

Marker-Inc-Korea/AutoRAG

Open-source framework for RAG evaluation and optimization via AutoML

GraphCanon updated 2w · GitHub synced 2w

5.0k stars419 forksLast push 2w TypeScript Apache-2.0

Decision brief

AutoRAG: Automate RAG task evaluation and optimization using AutoML techniques.

Good fit when

  • Automated benchmarking is needed for retrieval-augmented generation tasks
  • Desire to optimize Python pipeline with machine learning-driven automation

Avoid when

  • Requirements exceed capabilities of open-source tools
  • No need for RAG-specific optimization and evaluation features

Observed Jul 15, 2026 · Source: enrich:decision_facts

Verify the decision

Adoption

Package downloads where a registry match exists. GitHub stars (4,968) are secondary evidence.

npm downloads (30d)
316·npm downloads API·2w

Maintenance and security

Full trust report
Maintenance
Very active (2d since push)
As of 2w
Provenance
Not a fork · Organization account
As of 2w
Security (OSV)
No lockfile
As of 1mo

Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.

Install

npm install AutoRAG
npm

Similar tools

Same-category neighbours. No typed graph edges are catalogued for this tool yet.

Evidence and technical details

Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.

Overview

AutoRAG provides tools for automating retrieval-augmented generation tasks, focusing on evaluation and optimization with support for Python.

Capability facts

CLI
CLI entrypoint

Source: package.json:bin|scripts · Aug 8, 2026

MCP server
No MCP server detected

Source: repo_scan · Aug 8, 2026

Languages
typescript, javascript

Source: github.language+package.json · Aug 8, 2026

Categories

Tags

README

Installation

Published as @autorag/librarian (dist bundled with Bun, runtime Node ≥ 24 or Bun):

bun add @autorag/librarian          # library
bun install -g @autorag/librarian   # autorag CLI

---

## Quick Start

```typescript
import { AutoRAGAgent } from "@autorag/librarian";

const agent = new AutoRAGAgent({
  searchPaths: ["/path/to/documents"],
});

const response = await agent.searchDocuments("summarize the compliance requirements");
console.log(response.answer);
for (const result of response.results) {
  console.log(`[${result.number}] ${result.title} — ${result.summary}`);
}

// Mark which results were useful — AutoRAG remembers for next time
agent.recordFeedbackByNumbers(response.sessionId, [1, 3], [2]);

searchDocuments() runs the Pi agent loop — it searches, reads, consults memory, curates, and finalizes through the emit_autorag_results structured tool — then returns a typed SearchDocumentsResponse. The caller consumes the structured payload directly; no assistant text parsing.

For agents

This page has a .md twin and JSON over the API.

Was this helpful?

Anonymous feedback helps us improve pages and translations.