ChatTemplate class

Defines how to format chat messages for a specific model.

Each LLM family expects a different prompt structure. A ChatTemplate encodes the special tokens / delimiters that surround system, user, and assistant messages so the model generates coherent multi-turn replies.

The {text} placeholder in each template string is replaced with the actual message content at formatting time.

const zephyr = ChatTemplate(
  system: '<|system|>\n{text}</s>\n',
  user:   '<|user|>\n{text}</s>\n',
  assistant: '<|assistant|>\n{text}</s>\n',
);

Constructors

ChatTemplate({required String system, required String user, required String assistant, String systemDefault = 'You are a helpful assistant.'})

Creates a ChatTemplate with the given format strings.

const

Properties

assistant → String

Template wrapping each assistant message. Must contain {text}.

final

hashCode → int

The hash code for this object.

no setter, inherited

runtimeType → Type

A representation of the runtime type of the object.

no setter, inherited

system → String

Template wrapping the system prompt. Must contain {text}.

final

systemDefault → String

Fallback system prompt used when none is provided by the caller.

final

user → String

Template wrapping each user message. Must contain {text}.

final

Methods

format({String? systemPrompt, required List<({String role, String text})> messages}) → String

Renders a full prompt from a list of conversation messages.

noSuchMethod(Invocation invocation) → dynamic

Invoked when a nonexistent method or property is accessed.

inherited

toString() → String

A string representation of this object.

inherited

Operators

operator ==(Object other) → bool

The equality operator.

inherited