Quickstart: providers

attune-rag ships optional LLM provider adapters for Claude and Gemini. Each adapter lazy-imports its SDK, so the core package installs without requiring either. Use list_available() to confirm which providers are ready on your system, then call get_provider() to get a running instance.

from attune_rag.providers import list_available
print(list_available())
# Example output: ['claude', 'gemini']

Prerequisites

• attune-rag installed in your environment
• At least one provider SDK installed:
  • Claude: pip install attune-rag[claude]
  • Gemini: pip install attune-rag[gemini]
• A valid API key for the provider you want to use

Steps

1. Check which providers are available.

   from attune_rag.providers import list_available
   
   print(list_available())
   # ['claude']  ← only providers whose SDKs are importable appear here
   

2. Instantiate a provider. Pass the provider name to get_provider(). Supply your API key as a keyword argument; if you omit it, the provider reads from its default environment variable.

   from attune_rag.providers import get_provider
   
   provider = get_provider("claude", api_key="sk-ant-...")
   

   If you pass an unrecognised name, get_provider() raises ValueError: Unknown provider {...}. Known providers: {...}.

3. Generate a response.

   import asyncio
   
   response = asyncio.run(
       provider.generate("Explain retrieval-augmented generation in one sentence.")
   )
   print(response)
   # Retrieval-augmented generation combines a retrieval step with a language
   # model so answers are grounded in specific documents rather than only
   # parametric knowledge.
   

4. Generate a response with citations (Claude only). Pass a list of CitationDocument objects and a query to generate_with_citations(). The returned CitedResponse contains the answer text and a tuple of claim-level citations.

   from attune_rag.providers import get_provider
   from attune_rag.providers.base import CitationDocument
   
   provider = get_provider("claude", api_key="sk-ant-...")
   docs = [
       CitationDocument(title="RAG overview", text="RAG grounds LLM outputs in retrieved documents."),
   ]
   result = asyncio.run(
       provider.generate_with_citations(docs, query="What does RAG do?")
   )
   print(result.text)
   print(result.claim_citations)
   

Next

Read the LLMProvider protocol reference to learn how to implement your own adapter that works anywhere attune-rag accepts a provider.