Skip to Content
Getting StartedInstallation

Installation

Get the rag-forge CLI running on your machine.

Prerequisites

  • Node.js 20+ — for the CLI and MCP server
  • Python 3.11+ with uv  — for the RAG pipeline and evaluator
  • A vector store (Qdrant recommended for local dev)
  • An LLM provider API key (OpenAI, Anthropic, or a local model via Ollama)

Install the CLI

npm install -g @rag-forge/cli
# or
pnpm add -g @rag-forge/cli

Verify:

rag-forge --version

Install the Python runtime

The CLI delegates all pipeline and evaluation work to Python packages. rag-forge-core and rag-forge-observability are libraries (you import them); rag-forge-evaluator ships a rag-forge-eval CLI entry point.

The recommended setup creates a dedicated venv outside your project so the audit toolchain is isolated from your application’s dependencies:

# One-time setup
mkdir -p ~/.rag-forge && cd ~/.rag-forge
uv venv
source .venv/bin/activate    # macOS/Linux
# .venv\Scripts\activate     # Windows
 
uv pip install rag-forge-core rag-forge-evaluator rag-forge-observability

Verify:

rag-forge-eval --help
python -c "import rag_forge_core; print(rag_forge_core.__name__)"

rag-forge-core handles chunking, embedding, retrieval, and generation. rag-forge-evaluator handles RAGAS/DeepEval scoring and the RAG Maturity Model audit. rag-forge-observability provides OpenTelemetry tracing and drift detection primitives.

Note: The rag-forge npm CLI invokes the Python evaluator via subprocess, so it expects python and these packages to be on PATH. If you use a dedicated venv (recommended), make sure it’s activated when you run rag-forge audit.

Next steps

Quick Start