Skip to main content
Edition: Community (llm_providers) RAGSuite is model-agnostic within the providers below. Configure your API keys (or a local Ollama host) and pick chat / embedding models per project in Console › Model Configuration. See Bring your own keys. The console model picker uses these curated catalogs. Live API or Ollama discovery can add more chat models when keys or hosts are available. Azure OpenAI has no curated list — you use the deployment names from your own Azure resource.

OpenAI

Azure OpenAI

Use OpenAI models deployed in your own Azure OpenAI resource. Set it up in Model Configuration: After the endpoint and API key are set, use Refresh to load your chat and embedding deployments, or type the names manually.

Anthropic

Mistral

Google Gemini

Custom LLM / Ollama

Local or custom inference via Custom LLM / Ollama. Curated chat and embedding IDs, plus setup: Ollama & air-gapped.

Runtime embedding extras

These may work at runtime but are not curated picker entries:
  • OpenAI: text-embedding-ada-002
  • Gemini: text-embedding-004
  • Ollama: nomic-embed-text, all-minilm, and Jina EN variants
Prefer the curated embedding for each provider unless your deployment already depends on an extra.

Enterprise quality loop

Compare Models runs one query across models side by side and scores answers — Enterprise feature on your infrastructure.