ml-rag

Name: ml-rag
Availability: InStock
Author: mila

Build and query a retrieval-augmented generation pipeline from documents.

Updated Apr 6, 2026

Search Tags:#knowledge base #RAG #document indexing #document search #question answering

Ready to run

npx skills add https://github.com/milasaurus/compound-ml --skill ml-rag

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ml-rag
Source: https://github.com/milasaurus/compound-ml/tree/main/plugins/compound-ml/skills/ml-rag
Command: npx skills add https://github.com/milasaurus/compound-ml --skill ml-rag

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires sentence-transformers, chromadb, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill helps you create a searchable knowledge base and answer questions about a document collection without needing ML expertise.

Core Features & Use Cases

Build a Searchable Index: Index a collection of documents into a searchable vector store.
Query and Answer: Ask questions and get answers with source citations from the index.
Use Case: Build a RAG pipeline from a directory of documents to answer questions about the content.

Quick Start

Use the ml-rag skill to build a RAG pipeline from the directory 'documents'.

Frequently Asked Questions about ml-rag

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build a RAG pipeline from a directory of documents?▼

To build a RAG pipeline from documents, you index the collection into a searchable vector store, then query it to answer questions with source citations. This skill automates chunking, embedding, indexing, and retrieval.

What is retrieval-augmented generation and how does it help with document search?▼

Retrieval-augmented generation enhances document search by embedding chunks into a vector database and retrieving relevant context to answer questions with citations. It creates a searchable knowledge base without requiring ML expertise.

Do I need sentence-transformers and chromadb to query a document knowledge base?▼

Yes, you need sentence-transformers to generate embeddings and chromadb to manage the vector database. These dependencies handle the automatic chunking, indexing, and retrieval required for question answering.

Can I get source citations when asking questions about my indexed documents?▼

Yes, querying the document index returns answers with source citations. The retrieval pipeline fetches relevant document chunks from the vector store to ground the generated response in your source material.

What is the best way to index a collection of documents for question answering?▼

The best way to index documents for question answering is using an automated RAG pipeline. It handles text chunking, generates embeddings with sentence-transformers, and stores them in chromadb for retrieval.

Are there limitations when using a local vector database for RAG?▼

Using a local vector database like chromadb for RAG relies on sentence-transformers for embeddings, requiring local compute resources. Scaling to extremely large document collections may face memory and retrieval latency constraints.