neuralplus::AIResponse Struct Reference

Normalized response returned by both one provider round and the full tool loop. More...

#include <types.hpp>

Public Member Functions
	AIResponse (Message message)
	Creates a normalized response around one assistant message.

Public Attributes
Message	message
	Assistant message returned by the provider.
FinishReason	finish_reason {FinishReason::unknown}
	Normalized provider stop reason.
TokenUsage	usage
	Per-round usage from generate_once or cumulative usage from generate.
std::string	provider_request_id
	Provider request/response correlation identifier.
std::string	provider_model
	Exact model identifier reported by the provider.
std::size_t	model_rounds {1}
	Provider rounds completed by the public generate call.
std::size_t	tool_calls {0}
	Tool invocations completed by the public generate call.
bool	requires_continuation {false}
JsonValue	provider_metadata = JsonValue::object()
	Opaque provider response data available to advanced applications.

Detailed Description

Normalized response returned by both one provider round and the full tool loop.

Examples

anthropic_custom_config.cpp, anthropic_multimodal.cpp, gemini_custom_config.cpp, gemini_multimodal.cpp, openai_compatible_custom_config.cpp, openai_compatible_multimodal.cpp, openai_custom_config.cpp, openai_multimodal.cpp, and simple_session.cpp.

Member Data Documentation

requires_continuation

bool neuralplus::AIResponse::requires_continuation {false}

Requests another provider round after replaying this assistant message.

Provider adapters use this for protocol-level continuation such as Anthropic pause_turn; applications normally inspect only the final response returned by AIClient::generate.

---

The documentation for this struct was generated from the following file:

• include/neuralplus/types.hpp