# superlinear-ai/raglite

🥤 RAGLite is a Python toolkit for Retrieval-Augmented Generation (RAG) with DuckDB or PostgreSQL

Repository: https://github.com/superlinear-ai/raglite
Canonical: https://ross.abutalabs.com/products/raglite
Language: Python
License: MPL-2.0
License Family: copyleft
Topics: llm, markdown, pdf, rag, retrieval-augmented-generation, sqlite, vector-search, pgvector, postgres, postgresql, reranking, late-chunking, late-interaction, colbert, evals, query-adapter, chainlit, duckdb
Last push: 2026-08-17T07:19:07+00:00

## Health v2 (maintenance only)
Score: 85/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 98, release rhythm 84, longevity 58
- inputs: {"age_days": 814, "days_push": 16, "days_rel": 107, "gap_med": 5, "n_releases_24m": 18}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1198, forks 108 (observed 2026-08-28T04:03:57.689685+00:00)

## What it is
RAGLite is a Python toolkit for building Retrieval-Augmented Generation (RAG) pipelines backed by DuckDB or PostgreSQL, with hybrid keyword and vector search, reranking, and PDF-to-Markdown ingestion. It supports any LLM provider via LiteLLM, including local llama-cpp-python models, and uses lightweight permissive open-source dependencies.

## Use cases
- build a RAG chatbot over PDF documents
- add hybrid keyword and vector search to a Postgres app
- run retrieval-augmented generation with local LLMs
- chunk and embed documents with late chunking
- evaluate and rerank RAG retrieval quality
- query documents with adaptive retrieval

## When to choose
- you want a lightweight RAG toolkit without PyTorch or LangChain
- you prefer DuckDB or PostgreSQL as your vector store
- you need PDF ingestion, hybrid search, and reranking out of the box
- you want to run LLMs locally or swap providers via LiteLLM

## When to avoid
- you need a fully managed RAG service
- you require a heavy multi-agent orchestration framework
- your stack is not Python

## Facets
- artifact type: library
- maturity: active
- function: rag, search-engine, vector-database, llm-inference, pdf, nlp, chatbot
- domain: large-language-models, databases, pdf, developer-tools
- platform: python, windows
- tags: duckdb, postgresql, pgvector, hybrid-search, reranking, late-chunking, colbert, litellm, chainlit, sqlite, retrieval-augmented-generation, natural-language-processing, macos, linux, docker

## Member repositories
- superlinear-ai/raglite (main) score 85

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:57.689685+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 2026-08-30T06:21:11.697452+00:00, confidence not recorded.
  - readme: https://github.com/superlinear-ai/raglite (fetched 2026-08-28T04:03:57.689685+00:00, sha 37bb0214e78d)
  - registry_pypi: https://pypi.org/pypi/raglite/json (fetched 2026-08-29T12:28:44.465711+00:00, sha 391882640b56)
- Data as of 2026-08-30T08:39:29.467469+00:00.