Model interfaces
Send structured requests to generation and embedding models through a consistent API surface.
Features
Platform capabilities
Capabilities for developers connecting models, shaping prompts, retrieving context, evaluating responses, and managing AI application runtime behavior.
-
-
Prompt composition
Compose reusable prompts, organize versions, and pass variables between workflow steps.
-
Context retrieval
Retrieve relevant source passages and supply them as context for model requests.
-
Response evaluation
Run repeatable test cases to compare prompt versions and review response behavior.
-
Runtime configuration
Tune model settings, organize releases, and apply configuration across environments.
-
Request visibility
Inspect request traces, latency, errors, and token usage by application or model.