RAGLite

TypeScript

import { Document } from "raglite-toolkit";

const doc = new Document("./policy.pdf", { llm: { provider: "anthropic" } });
await doc.build();

const answer = await doc.ask("What is the refund policy?");
console.log(answer.text, answer.citations);

Python

from raglite import Document

doc = Document("./policy.pdf", {"llm": {"provider": "anthropic"}})
doc.build()

answer = doc.ask("What is the refund policy?")
print(answer.text, answer.citations)

Answers with citations

Every answer lists the passages it used, with the PDF page and Markdown section of each. Citations

Hybrid search

Vector search for meaning, BM25 keyword search for exact terms such as error codes, or both fused. Hybrid search

MCP server and agent tools

Give Claude Code, Cursor or VS Code your documents with one command, or hand a search tool to your own agent. MCP server

Any provider

OpenAI, Anthropic, Google, Mistral, Cohere, Groq, xAI and Ollama. Install only the SDKs you use. Providers

Fully local

Local embeddings and Ollama answers: no API key, nothing leaves your machine. Getting started

Two SDKs that match

The TypeScript and Python packages have the same features, options, tools and REST API, and shared test fixtures keep them in step.