Function Calling

Let on-device models call external functions and integrate with other services.

Function calling lets a model request that your app run an external function — for example, changing the UI, querying a database, or calling another service — and then continue the conversation with the result.

Supported models#

Models with function calling support#

Gemma 4 (E2B, E4B) — full support (native function-call tokens).
Gemma3n (E2B, E4B) — full support.
Gemma 3 1B — function calling support.
FunctionGemma 270M — Google's specialized function-calling model.
DeepSeek R1 — function calling + thinking mode.
Qwen models (0.5B, 0.6B, 1.5B) — full support.
Phi-4 Mini — advanced reasoning with function calling.

Models without function calling support#

Gemma 3 270M — text generation only.
SmolLM 135M — text generation only.
FastVLM 0.5B — vision model, no function calling.

When you pass tools to an unsupported model, the plugin logs a warning and ignores the tools — the model still works normally for text generation. Check the `supportsFunctionCalls` property in your model configuration.

Handling function calls#

When the model wants to call a function, the response stream emits a FunctionCallResponse with the function name and arguments. Execute it, then send a Message.toolResponse(...) back to the model:

chat.generateChatResponseAsync().listen((response) {
  if (response is TextResponse) {
    // Regular text token
    print('Text token: ${response.token}');
  } else if (response is FunctionCallResponse) {
    // Model wants to call a function
    print('Function: ${response.name}');
    print('Arguments: ${response.args}');
    _handleFunctionCall(response);
  }
});

Send the function result back to the model so it can continue:

final toolMessage = Message.toolResponse(
  toolName: 'change_background_color',
  response: {'status': 'success', 'color': 'blue'},
);
await chat.addQueryChunk(toolMessage);
final followUp = await chat.generateChatResponse();

Platform support#

Function calling is supported on Android, iOS, Web, and Desktop. For Gemma 4, the native function-call tokens are routed through the LiteRT-LM SDK chat-template path (use ModelType.gemma4).

Function calling / tool calls are **not** supported on the web `.litertlm` path (`@litert-lm/core` early preview). For function calling on web, use MediaPipe `.task` web models. See [Troubleshooting](/docs/troubleshooting).

See Models for the correct ModelType per model family.