const core::Array<ChatMessage>& transcript,

const core::Array<ToolDefinition>& tools) const

{

const OpenAICompatOutboundRequest request = build_openai_chat_request(config_, transcript, tools);

ProviderRequest result;

result.valid = request.valid;

result.url = request.url;

result.body = request.body;

result.error_message = request.error_message;

// No auth header for a local Ollama peer - result.headers stays empty.

return result;

const int32_t status_code,

const core::String& response_body,

const bool transport_ok) const

{

const OpenAICompatChatResponse response = parse_openai_chat_response(status_code, response_body, transport_ok);

ProviderResponse result;

result.transport_ok = response.transport_ok;

result.http_success = response.http_success;

result.assistant_message = response.assistant_message;

result.thinking_text = response.thinking_text;

result.tool_calls = response.tool_calls;

result.error_message = response.error_message;

result.prompt_tokens = response.prompt_tokens;

result.completion_tokens = response.completion_tokens;

result.done_reason = response.done_reason;

return result;