Skip to main content

AI Providers

AI Providers hold the credentials and connection details for a large language model service (OpenAI, Anthropic, Azure OpenAI, Ollama, a custom OpenAI-compatible endpoint, or Amazon Bedrock), or for TypeSafe, whose Jev model grades with calibrated probabilities instead of generated text.

Secret fields are encrypted at rest. Providers are referenced by AI Judges and can be tested against a caller-specified model.