For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Python class

TextGenerationResponseFormat

`TextGenerationResponseFormat`

class max.pipelines.context.TextGenerationResponseFormat(type, json_schema=<factory>, grammar=None, grammar_enforced=False, tools_forced=False, requires_structured_output_flag=False, has_json_schema=False)

source

Bases: object

Represents the response format specification for a text generation request.

Parameters:

type (str)
json_schema (dict[str, Any])
grammar (str | None)
grammar_enforced (bool)
tools_forced (bool)
requires_structured_output_flag (bool)
has_json_schema (bool)

`grammar`

grammar: str | None = None

source

Grammar for constrained decoding.

When set with type="grammar", this takes precedence over json_schema. Used for model-specific constrained decoding formats like Kimi’s tool call grammar.

`grammar_enforced`

grammar_enforced: bool = False

source

Whether to actively enforce grammar via bitmask.

When True from the start, enforce grammar from the first token. When False initially (for tool_choice=auto without response_format), the grammar is compiled but not enforced until a tool call start token is detected.

`has_json_schema`

has_json_schema: bool = False

source

Whether this request includes a JSON schema response format.

`json_schema`

json_schema: dict[str, Any]

source

A JSON schema dictionary that defines the structure and validation rules for the generated response.

`requires_structured_output_flag`

requires_structured_output_flag: bool = False

source

Whether this request requires --enable-structured-output to be set.

True when the constraint includes a user-supplied JSON schema (from response_format). False for pure tool-call grammars derived from the model’s tool parser, which work without the operator flag because the grammar is server-controlled, not user-controlled.

`tools_forced`

tools_forced: bool = False

source

Whether tool calling was forced (tool_choice=required or named function).

Controls whether grammar_enforced is True from the first generated token. Independent of the --enable-structured-output flag (which only gates user-supplied schemas; see requires_structured_output_flag).

`type`

type: str

source

The type of response format, for example, json_object or grammar.

TextGenerationResponseFormat​

grammar​

grammar_enforced​

has_json_schema​

json_schema​

requires_structured_output_flag​

tools_forced​

type​