Fifty Speech Engine #

One engine, two speech modes -- TTS narration and STT recognition with builder-customizable controls.

Unifies flutter_tts and speech_to_text behind a single initialize()/speak()/startListening() API. Ships three FDL-styled control widgets (SpeechTtsControls, SpeechSttControls, SpeechControlsPanel) with optional contentBuilder callbacks that let you replace the default UI while keeping all state management intact. Part of Fifty Flutter Kit.

TTS Panel	STT Panel

Why fifty_speech_engine #

One engine, two speech modes -- FiftySpeechEngine unifies flutter_tts and speech_to_text behind a single initialize()/speak()/startListening() API; no managing two SDKs.
Custom TTS and STT controls, logic unchanged -- SpeechTtsControls and SpeechSttControls accept optional contentBuilder callbacks; replace the default FDL UI while keeping all state management intact.
Type-safe builder state -- Builders receive SpeechTtsState (enabled, rate, pitch, volume, isSpeaking) and SpeechSttState (enabled, isListening, recognizedText, isAvailable) data classes, not raw booleans; your custom UI compiles cleanly.
Continuous and command modes -- Set listenContinuously: true for dictation, false for single voice commands; the same engine handles both.

Installation #

dependencies:
  fifty_speech_engine: ^0.2.1

For Contributors #

dependencies:
  fifty_speech_engine:
    path: ../fifty_speech_engine

Dependencies: speech_to_text, flutter_tts, fifty_tokens, fifty_theme, fifty_ui

Quick Start #

import 'package:fifty_speech_engine/fifty_speech_engine.dart';

// 1. Create engine with locale
final engine = FiftySpeechEngine(
  locale: Locale('en', 'US'),
  onSttResult: (text) async => print('Heard: $text'),
  onSttError: (error) => print('Error: $error'),
);

// 2. Initialize
await engine.initialize();

// 3. Speak text
await engine.speak('Hello! How can I help you?');

// 4. Listen for voice input
await engine.startListening(continuous: true);

// 5. Stop listening when done
await engine.stopListening();

// 6. Clean up
engine.dispose();

Architecture #

FiftySpeechEngine
    |
    +-- tts: TtsManager
    |       Text-to-Speech synthesis via flutter_tts
    |
    +-- stt: SttManager
            Speech-to-Text recognition via speech_to_text

Core Components #

Component	Description
`FiftySpeechEngine`	Unified interface combining TTS and STT
`TtsManager`	Text-to-Speech handler with voice configuration
`SttManager`	Speech-to-Text handler with queue support
`SpeechResultModel`	Result container for recognized speech

Widgets #

Three FDL-styled control widgets for TTS and STT. All are callback-based (no internal state management dependency) and accept optional builder callbacks for custom UI.

SpeechTtsControls #

TTS enable/disable toggle with optional rate, pitch, and volume sliders. Speaking indicator shown when active.

SpeechTtsControls(
  enabled: ttsEnabled,
  onEnabledChanged: (value) => setState(() => ttsEnabled = value),
  rate: rate,
  onRateChanged: (value) => setState(() => rate = value),
  pitch: pitch,
  onPitchChanged: (value) => setState(() => pitch = value),
  volume: volume,
  onVolumeChanged: (value) => setState(() => volume = value),
  isSpeaking: isSpeaking,
)

Parameter	Type	Default	Description
`enabled`	`bool`	required	Whether TTS is enabled
`onEnabledChanged`	`ValueChanged<bool>`	required	Toggle callback
`rate`	`double`	`1.0`	Speech rate (0.5--2.0)
`onRateChanged`	`ValueChanged<double>?`	`null`	Rate slider callback; null hides the slider
`pitch`	`double`	`1.0`	Speech pitch (0.5--2.0)
`onPitchChanged`	`ValueChanged<double>?`	`null`	Pitch slider callback; null hides the slider
`volume`	`double`	`1.0`	Speech volume (0.0--1.0)
`onVolumeChanged`	`ValueChanged<double>?`	`null`	Volume slider callback; null hides the slider
`isSpeaking`	`bool`	`false`	Whether TTS is currently speaking
`compact`	`bool`	`false`	Use compact layout
`showCard`	`bool`	`true`	Wrap in FiftyCard; set false when embedding
`contentBuilder`	`SpeechTtsContentBuilder?`	`null`	Custom content builder (see Customization)

SpeechSttControls #

STT enable/disable toggle with microphone button, pulsing listening indicator, recognized text display, and error handling.

SpeechSttControls(
  enabled: sttEnabled,
  onEnabledChanged: (value) => setState(() => sttEnabled = value),
  isListening: isListening,
  onListenPressed: toggleListening,
  recognizedText: recognizedText,
  onClear: () => setState(() => recognizedText = ''),
  errorMessage: errorMessage,
)

Parameter	Type	Default	Description
`enabled`	`bool`	required	Whether STT is enabled
`onEnabledChanged`	`ValueChanged<bool>`	required	Toggle callback
`isListening`	`bool`	required	Whether STT is currently listening
`onListenPressed`	`VoidCallback`	required	Microphone button callback
`recognizedText`	`String`	`''`	Last recognized text
`isAvailable`	`bool`	`true`	Whether STT is available on this device
`errorMessage`	`String?`	`null`	Error message to display
`onClear`	`VoidCallback?`	`null`	Clear text callback; null hides the clear button
`compact`	`bool`	`false`	Use compact layout
`showCard`	`bool`	`true`	Wrap in FiftyCard; set false when embedding
`hintText`	`String?`	`'TAP TO SPEAK'`	Hint text below microphone button
`contentBuilder`	`SpeechSttContentBuilder?`	`null`	Custom content builder (see Customization)

SpeechControlsPanel #

Combined panel with both TTS and STT controls in a single card. Supports showTts and showStt flags to selectively display either section.

SpeechControlsPanel(
  ttsEnabled: ttsEnabled,
  onTtsEnabledChanged: (v) => setState(() => ttsEnabled = v),
  rate: rate,
  onRateChanged: (v) => setState(() => rate = v),
  isSpeaking: isSpeaking,

  sttEnabled: sttEnabled,
  onSttEnabledChanged: (v) => setState(() => sttEnabled = v),
  isListening: isListening,
  onListenPressed: toggleListening,
  recognizedText: recognizedText,
)

Parameter	Type	Default	Description
`showTts`	`bool`	`true`	Show TTS controls section
`showStt`	`bool`	`true`	Show STT controls section
`compact`	`bool`	`false`	Use compact layout for both sections
`title`	`String?`	`null`	Optional panel title
`ttsBuilder`	`SpeechTtsContentBuilder?`	`null`	Custom TTS content builder
`sttBuilder`	`SpeechSttContentBuilder?`	`null`	Custom STT content builder

All TTS and STT parameters from the individual widgets are also available on the panel.

Customization #

All three widgets accept optional builder callbacks. When provided, the builder replaces the default FDL rendering while the widget retains ownership of the FiftyCard wrapper (when showCard is true) and all state computation.

Custom TTS Controls #

Replace the default TTS controls layout. The builder receives a SpeechTtsState data class with all state values and callbacks:

SpeechTtsControls(
  enabled: ttsEnabled,
  onEnabledChanged: (v) => setState(() => ttsEnabled = v),
  rate: rate,
  onRateChanged: (v) => setState(() => rate = v),
  isSpeaking: isSpeaking,
  contentBuilder: (state) {
    return Column(
      children: [
        Switch(value: state.enabled, onChanged: state.onEnabledChanged),
        if (state.enabled) ...[
          Slider(value: state.rate, min: 0.5, max: 2.0, onChanged: state.onRateChanged),
          if (state.isSpeaking) const Text('Speaking...'),
        ],
      ],
    );
  },
)

SpeechTtsState

Field	Type	Description
`enabled`	`bool`	Whether TTS is enabled
`onEnabledChanged`	`ValueChanged<bool>`	Toggle callback
`rate`	`double`	Current speech rate
`onRateChanged`	`ValueChanged<double>?`	Rate change callback
`pitch`	`double`	Current pitch
`onPitchChanged`	`ValueChanged<double>?`	Pitch change callback
`volume`	`double`	Current volume
`onVolumeChanged`	`ValueChanged<double>?`	Volume change callback
`isSpeaking`	`bool`	Whether TTS is speaking
`compact`	`bool`	Whether compact mode is active

Custom STT Controls #

Replace the default STT controls layout. The builder receives a SpeechSttState data class:

SpeechSttControls(
  enabled: sttEnabled,
  onEnabledChanged: (v) => setState(() => sttEnabled = v),
  isListening: isListening,
  onListenPressed: toggleListening,
  recognizedText: recognizedText,
  contentBuilder: (state) {
    return Column(
      children: [
        Switch(value: state.enabled, onChanged: state.onEnabledChanged),
        if (state.enabled) ...[
          IconButton(
            icon: Icon(state.isListening ? Icons.mic : Icons.mic_none),
            onPressed: state.onListenPressed,
          ),
          if (state.recognizedText.isNotEmpty)
            Text('"${state.recognizedText}"'),
        ],
      ],
    );
  },
)

SpeechSttState

Field	Type	Description
`enabled`	`bool`	Whether STT is enabled
`onEnabledChanged`	`ValueChanged<bool>`	Toggle callback
`isListening`	`bool`	Whether STT is listening
`onListenPressed`	`VoidCallback`	Microphone button callback
`recognizedText`	`String`	Last recognized text
`isAvailable`	`bool`	Whether STT is available
`errorMessage`	`String?`	Error message
`onClear`	`VoidCallback?`	Clear text callback
`compact`	`bool`	Whether compact mode is active
`hintText`	`String?`	Hint text for microphone button

Custom Panel Sections #

SpeechControlsPanel forwards builder callbacks to the internal SpeechTtsControls and SpeechSttControls:

SpeechControlsPanel(
  ttsEnabled: ttsEnabled,
  onTtsEnabledChanged: (v) => setState(() => ttsEnabled = v),
  isSpeaking: isSpeaking,

  sttEnabled: sttEnabled,
  onSttEnabledChanged: (v) => setState(() => sttEnabled = v),
  isListening: isListening,
  onListenPressed: toggleListening,
  recognizedText: recognizedText,

  ttsBuilder: (state) => MyTtsControls(state: state),
  sttBuilder: (state) => MySttControls(state: state),
)

Builder Signatures #

Widget	Builder parameter	Callback signature
`SpeechTtsControls`	`contentBuilder`	`Widget Function(SpeechTtsState state)`
`SpeechSttControls`	`contentBuilder`	`Widget Function(SpeechSttState state)`
`SpeechControlsPanel`	`ttsBuilder`	`Widget Function(SpeechTtsState state)`
`SpeechControlsPanel`	`sttBuilder`	`Widget Function(SpeechSttState state)`

All builders are optional. Omit them to use the default FDL UI.

API Reference #