Expand description
Message and request/response models.
Structs§
- Grounding
Chunk - A source chunk used for grounding.
- Grounding
Metadata - Grounding and citation metadata.
- Grounding
Support - Links a segment of generated content to grounding sources.
- Message
- Conversation message with role and content blocks.
- Message
Options - Options for LLM message requests.
- Metadata
- Metadata for request tracking.
- Response
Metadata - Provider-specific metadata returned with responses.
- Safety
Rating - Content safety rating from the model.
- Tool
- Tool definition for function calling.
- Tool
Result - Result of a tool invocation, sent back to the assistant.
- ToolUse
- Tool invocation requested by the assistant.
- Usage
- Token usage statistics.
Enums§
- Content
- Content block types that can appear in messages.
- Content
Block Type - Type of content block in streaming.
- Image
Source - Image source types supported by LLM providers.
- Role
- Message role in a conversation.
- Stream
Event - Events emitted during streaming responses.
- Tool
Choice - Controls how the model uses tools.