chat_out.hpp File Reference

Chat Output Parsing. More...

#include "common.hpp"
#include <llama/llama.h>
#include <chat.h>
#include <peg-parser.h>
#include <exception>
#include <string>
#include <vector>

Go to the source code of this file.

Classes
struct	lloyal::chat_out::ToolCall
	A single tool call extracted from model output. More...
struct	lloyal::chat_out::ParseResult
	Result from parsing model output. More...

Namespaces
namespace	lloyal
	Boundary Tracker Stub for OSS liblloyal.
namespace	lloyal::chat_out
	Chat output parsing (tool calls, reasoning, content)

Functions
ParseResult	lloyal::chat_out::parse (const std::string &output, common_chat_format format, common_reasoning_format reasoning_format=COMMON_REASONING_FORMAT_NONE, bool is_partial=false, bool thinking_forced_open=false, const std::string &parser_data="")
	Parse model output with explicit format.
ParseResult	lloyal::chat_out::parse (const llama_model *model, const std::string &output, bool is_partial=false)
	Parse model output with auto-detected format from model template.

Detailed Description

Chat Output Parsing.

Wraps llama.cpp's common_chat_parse() to extract structured content from model output: plain text, reasoning/thinking content, and tool calls. Stateless — each call is independent.

Usage Pattern

Typically paired with chat_in::format():

1. chat_in::format() → FormatResult (includes format enum + reasoning_format)
2. Generate tokens using the formatted prompt
3. chat_out::parse() with the format from step 1 → ParseResult

Dependencies

• common/chat.h: common_chat_parse()

See also

lloyal::chat_in for input formatting

Definition in file chat_out.hpp.