🦙 TryToSpeak
AI Language Assistant

🦙 TryToSpeak
AI Language Assistant

Loading...
Loading Authentication...

👋 , !

Model:
Models Available:
Port:
🎛️ Chat Settings
🤖 Model Parameters
Maximum response length
Response creativity (0=focused, 2=creative)
How the model processes responses
💭 Conversation History Settings
Number of past exchanges to include as context
Save conversation history to disk per model
Automatically open microphone after audio response ends
🖥️ Interface Display Settings
Display the message input box and send controls
Display the model response box
Switch to dark theme for better viewing in low light
🎵 Audio & Language Settings
Response language and voice selection
Generate spoken audio responses
Choose preferred voice for TTS
Current:
Adjust voice playback speed

🎵 Available Voices for

📚 Available Models ADMIN ONLY

Loading models...

🔽 Download New Model ADMIN ONLY

🔌 API Endpoints Documentation

GET /health

Description: Check server health and status

Response: JSON with server status, model info, memory usage, and availability

curl -X GET "http://localhost:5002/health"
{ "status": "healthy", "llamacpp_available": true, "current_model": "Qwen3-Zro-Cdr-Reason-V2-0.8B-NEO-EX-D_AU-Q4_K_M-imat.gguf", "models_count": 5, "memory_status": { "total_ram_gb": 16.0, "available_ram_gb": 8.5, "used_percent": 47.2, "safe_context_length": 4096 } }
POST /api/chat

Description: Generate text response with advanced parameters and thinking modes. When TTS is enabled, provides precise audio duration information for timing and synchronization.

Body: JSON or form data with message and optional parameters

curl -X POST "http://localhost:5002/api/chat" \ -H "Content-Type: application/json" \ -d '{ "message": "What is consciousness?", "max_tokens": 300, "temperature": 0.7, "thinking_mode": "thinking-full", "language": "english", "include_audio": true, "voice": "en-US-JennyNeural", "voice_speed": 1.0, "system_instructions": "You are a psychology expert.", "context_messages": 50, "preserve_history": true }'
{ "success": true, "response": "Consciousness is a fascinating topic...", "model": "lucy_128k-Q3_K_M.gguf", "audio_url": "/audio", "audio_duration_seconds": 31.2, "audio_size_kb": 486.7, "voice_speed": 1.0 }

⏱️ Audio Duration & Timing Features

Precise Duration Calculation: The server uses mutagen library to extract exact audio duration from generated MP3 files for accurate timing information.

Integration Use Cases: Voice command systems, progress indicators, mobile apps, batch processing, and accessibility features.