| --- |
| title: VoiceCal - Voice-Enabled AI Scheduling Assistant |
| emoji: 🎤📅 |
| colorFrom: blue |
| colorTo: purple |
| sdk: streamlit |
| app_file: app.py |
| pinned: false |
| license: mit |
| short_description: Voice-enabled AI assistant with WebRTC integration |
| --- |
| |
| # 🎤📅 VoiceCal - Voice-Enabled AI Scheduling Assistant |
|
|
| An intelligent AI scheduling assistant with **voice interaction capabilities** powered by WebRTC, Whisper STT, and Bark TTS. Book Google Calendar appointments through natural conversation - now with voice input and audio responses! |
|
|
| > **🚀 New Deployment**: VoiceCal is the latest iteration with improved WebSocket-based voice services and enhanced reliability. Previous ChatCal.ai space has been superseded by this deployment. |
|
|
| ## 🎯 Features |
|
|
| ### 🗣️ Voice Interaction |
| - 🎤 **Real-time Speech-to-Text**: WebRTC audio capture with Whisper transcription |
| - 🔊 **Text-to-Speech Responses**: AI responses with natural voice synthesis |
| - 🎭 **Multiple Voice Options**: Choose from different voice personalities |
| - ⚡ **Real-time Processing**: Live transcription as you speak |
|
|
| ### 📅 Smart Calendar Integration |
| - 🤖 **AI-Powered Booking**: Natural language appointment scheduling |
| - 📅 **Google Calendar Sync**: Seamless integration with your calendar |
| - 🔍 **Conflict Detection**: Smart availability checking |
| - 🎥 **Google Meet Integration**: Automatic video conference setup |
| - 📧 **Email Notifications**: Booking confirmations and cancellations |
|
|
| ### 🧠 Intelligent Conversation |
| - 💭 **Conversation Memory**: Persistent context across interactions |
| - 🎯 **Smart Extraction**: Automatically extract names, emails, times |
| - 🌍 **Timezone Awareness**: Global scheduling support |
| - ⏰ **Flexible Time Parsing**: "tomorrow at 2pm", "next Tuesday", etc. |
|
|
| ## 🏗️ Architecture |
|
|
| - **Frontend**: Gradio with WebRTC audio capture |
| - **AI**: Groq Llama-3.1 with Anthropic Claude fallback |
| - **STT**: Whisper via external service integration |
| - **TTS**: Bark text-to-speech synthesis |
| - **Calendar**: Google Calendar API with OAuth2 |
| - **Storage**: Google Cloud Secret Manager for persistent auth |
|
|
| ## 🚀 Usage |
|
|
| ### Voice Interaction |
| 1. Click the microphone button to start recording |
| 2. Speak naturally: "Hi, I'm John. Book a 30-minute meeting tomorrow at 2pm" |
| 3. Watch real-time transcription appear in the text box |
| 4. AI responds with voice confirmation of your booking |
|
|
| ### Text Interaction |
| - Type messages as normal - voice and text work together |
| - Edit voice transcriptions before sending |
| - Use quick action buttons for common requests |
|
|
| ### Example Conversations |
|
|
| **Voice**: "Book a Google Meet with Peter next Tuesday at 10 AM for 45 minutes" |
| **AI Audio Response**: "Perfect! I've scheduled your 45-minute Google Meet with Peter for next Tuesday at 10:00 AM..." |
|
|
| ## 🛠️ Development |
|
|
| This space integrates: |
| - **VoiceCal Core**: Calendar booking logic and Google integration |
| - **STT Service**: External Whisper service for speech recognition via WebSocket |
| - **TTS Service**: External Bark service for voice synthesis via WebSocket |
| - **WebRTC**: Browser-based audio capture and streaming |
|
|
| ### 🔧 Service Configuration |
|
|
| For optimal performance, the external voice services should be configured with: |
| - **STT Service**: `STT_SERVICE_MODE=websocket` (enables WebSocket endpoints) |
| - **TTS Service**: `TTS_SERVICE_MODE=websocket` (enables WebSocket endpoints) |
| - **Service URLs**: |
| - STT: `wss://pgits-stt-gpu-service.hf.space/ws/stt` |
| - TTS: `wss://pgits-tts-gpu-service.hf.space/ws/tts` |
|
|
| ## 🔐 Privacy & Security |
|
|
| - Secure OAuth2 authentication with Google |
| - Audio processed in real-time, not stored |
| - Persistent token storage via Google Secret Manager |
| - All calendar operations respect your existing permissions |
|
|
| ## 📞 Contact |
|
|
| For business scheduling needs or technical support: pgits.job@gmail.com |
|
|
