Configuration - Ragrails

Ragrails reads credentials from environment variables. Use a .env file at the project root or export them in your shell.

Embedding providers

Provider	Variable
Voyage	`VOYAGE_API_KEY`

Provider	Variable	Models
OpenAI	`OPENAI_API_KEY`	`gpt-5.5`, `gpt-5.4-mini`, `gpt-4.1`, `gpt-4o`, `gpt-4o-mini`, and more
Anthropic	`ANTHROPIC_API_KEY`	`claude-opus-4-8`, `claude-sonnet-4-6`, `claude-haiku-4-5`, and more
Google	`GEMINI_API_KEY` or `GOOGLE_API_KEY`	`gemini-3-pro`, `gemini-2.5-pro`, `gemini-2.5-flash`, `gemini-2.5-flash-lite`, and more

Google Gemini models don’t support tool calling yet, so they can’t be used with the interactive CLI’s agentic tools.

Provider	Variable
Voyage	`VOYAGE_API_KEY`
BM25 (local)	none, runs offline

Provider	Variables
Qdrant (local)	`QDRANT_URL` (default: `http://localhost:6333`)
Qdrant (cloud)	`QDRANT_URL`, `QDRANT_API_KEY`
Pinecone	`PINECONE_API_KEY`
Weaviate	`WEAVIATE_URL`, `WEAVIATE_API_KEY`

from dotenv import load_dotenv
load_dotenv()

from ragrails import RagRails
rag = RagRails()

URL ingestion uses Playwright via crawl4ai. Install the extra and run browser setup once per environment:

pip install "ragrails[url]"

RagRails().setup_url()              # installs chromium (default)
RagRails().setup_url(browser="firefox")

Or from the CLI:

ragrails setup-url
ragrails setup-url --browser firefox