Skip to main content
The Settings panel allows you to customize ChatbotAI-Free’s behavior and appearance. Access it by clicking the ⚙️ Settings button in the main interface.

​
Language Selection

Choose your preferred language for both speech recognition and text-to-speech:
  • English - Uses Whisper STT with en language code
  • Español (Spanish) - Uses Whisper STT with es language code
Changing the language affects both speech-to-text transcription and the available voice options. Kokoro TTS supports both English and Spanish natively.
The language setting is stored in user_preferences.json:36 and defaults to "english".

​
Voice Speed Adjustment

Control how fast the AI speaks with the voice speed multiplier:
1

Range

Adjust speed from 0.5× to 2.0× using the slider or input field
2

Default

The default speed is 1.0× (normal speaking pace)
3

Apply

Changes take effect immediately for the next TTS response
The voice speed setting is passed to both Kokoro and Sherpa-ONNX TTS engines via the speed parameter in tts_manager.py:94.

​
Font Size Options

Choose from three preset font sizes to improve readability:
SizeBubble TextInput TextStatus Text
Small13px12px11px
Medium (default)15px14px13px
Large18px16px15px
Font size configurations are defined in preferences.py:42-58.

​
Recording Mode Settings

Configure how voice recording behaves:

​
Silence Duration

The recorder uses Voice Activity Detection (VAD) to automatically stop recording after detecting silence:
  • Default: 3.0 seconds of silence
  • Adjustable: Change in audio_utils.py:27
  • Minimum audio duration: 1.0 seconds (filters out clicks and noise)
The silence threshold is set to 0.03 RMS to avoid triggering on background noise. This is configured in audio_utils.py:21.

​
Sample Rate

The recorder uses a 16000 Hz sample rate optimized for Whisper STT. If your microphone’s native sample rate differs, audio is automatically resampled (see audio_utils.py:61-63).

​
Auto-Send Configuration

Control whether recorded voice messages are sent automatically:
When enabled (auto_send: true), voice recordings are automatically transcribed and sent to the AI immediately after silence is detected.Use case: Hands-free conversation in Live Mode
When disabled (auto_send: false), you must manually press the send button after recording.Use case: Review transcriptions before sending in Classic Chat mode
The auto-send preference is stored in preferences.py:30 and defaults to true.

​
Saving Settings

All settings are automatically saved to user_preferences.json when you click the Save or Apply button. The preferences file is located in the application’s root directory.
Changes to the Whisper model size require an application restart to take effect. You’ll be prompted to restart immediately after changing this setting.

​
Default Settings

If the preferences file doesn’t exist or is corrupted, the app uses these defaults (from preferences.py:28-40): 
{
  "model": null,
  "auto_send": true,
  "font_size": "medium",
  "language": "english",
  "voice_speed": 1.0,
  "english_voice": "af_bella",
  "spanish_voice": "ef_dora",
  "output_device": -1,
  "input_device": -1,
  "context_size": 0,
  "whisper_model": "base"
}

Build docs developers (and LLMs) love

Get started for freeTalk to us