RAG-Forge

Framework-agnostic CLI toolkit for production-grade RAG pipelines with evaluation baked in.

RAG-Forge gives you the building blocks for retrieval-augmented generation — ingestion, chunking, hybrid retrieval, evaluation, drift detection, cost tracking, and observability — wired together with a single CLI. No framework lock-in. Bring your own LLM provider, your own vector store, your own embeddings.

Start here

• Installation — Get the CLI running in under a minute.
• Quick Start — Index a corpus and run your first query.
• Core concepts — The mental model before you build.

Reference

• CLI Reference — Every rag-forge command, every flag.
• Templates — Scaffolds for common RAG patterns.
• Concepts — How chunking, retrieval, and evaluation work under the hood.
• MCP Server — Expose RAG-Forge to Claude Desktop and other MCP clients.

Why RAG-Forge

RAG is hard to get right. Most teams ship a prototype, watch it break on real queries, and have no framework for diagnosing why. RAG-Forge is built around the RAG Maturity Model (RMM) — a 0-to-5 scale that turns vague “our RAG is broken” conversations into specific, measurable improvements. Every command is designed to move you up one level.