Using Vaani#
How to interact with Vaani and use its capabilities.
First Interaction#
Start Vaani
python3 main.py
You’ll see:
Vaani Assistant starting...
Listening...
[Microphone icon]
Say Something
Speak naturally to the microphone. Say:
“Hey Aria, what time is it?”
“What’s the weather like?”
“Play some music”
Vaani will:
Recognize your speech
Think about your request
Speak a response
Play audio/music if needed
Common First Commands
"What's your name?"
→ Vaani tells you about itself
"What can you do?"
→ Lists capabilities
"Tell me about yourself"
→ Personality introduction
"Play some rock music"
→ Finds and plays YouTube music
"What's the weather in London?"
→ Searches and reports weather
Core Commands#
Information Queries
Ask Vaani questions and it searches the web for answers:
"What's the capital of France?"
"How tall is Mount Everest?"
"When was the internet invented?"
"Who won the World Cup?"
"What's the temperature outside?"
Vaani will search the web and provide current information.
Time and Reminders
"What time is it?"
→ Current time in your timezone
"What's today's date?"
→ Today's date
"Set a reminder for lunch"
→ Sets a reminder (supported in some configurations)
Music Playback
Vaani can find and play music from YouTube:
"Play some jazz"
→ Random jazz playlist
"Play [Artist Name] [Song Name]"
→ Specific song, e.g. "Play The Beatles Let It Be"
"Play rock music"
→ Random rock songs
"Play movie soundtracks"
→ Instrumental/theme music
Music plays through your system audio.
Conversation
Vaani remembers conversation context:
You: "What's the weather in New York?"
Aria: "It's 72°F and sunny"
You: "How about Tokyo?"
Aria: "In Tokyo it's 28°C and partly cloudy"
Vaani understands "Tokyo" refers to your previous question
about weather, not a separate query.
About Vaani
"Tell me about yourself"
→ Vaani's background and purpose
"What's your personality like?"
→ Description of Vaani's communication style
"How do you work?"
→ Technical explanation of capabilities
Advanced Usage#
Multi-Turn Conversations
Vaani maintains conversation context across turns:
You: "Who's the president of the United States?"
Aria: "As of 2024, it's Joe Biden"
You: "How old is he?"
Aria: "He was born on November 20, 1942, making him 81 years old"
You: "What about his vice president?"
Aria: "That would be Kamala Harris"
Each response uses the previous context.
Fact-Based Queries
For questions requiring current information, Vaani searches the web:
"What's trending on Twitter?"
"What are the latest tech news?"
"How's the stock market doing?"
"What's happening in Ukraine?"
Specific Artist/Song Requests
For exact songs, use the full song name:
✓ "Play Hotel California by the Eagles"
✓ "Play Bohemian Rhapsody Queen"
✓ "Play Stairway to Heaven"
✗ "Play that song from the 80s" (too vague)
✗ "Play something I heard on the radio" (too vague)
Conversation About Results
After Vaani provides information, discuss it:
You: "What's Python?"
Aria: "Python is a programming language known for..."
You: "What are its main uses?"
Aria: "The main uses are web development, data science..."
You: "Can I learn it quickly?"
Aria: "Yes, Python is known for being beginner-friendly..."
Clarifying Follow-ups
If Vaani misunderstands:
You: "Tell me about Java"
Aria: "Java is an island in Indonesia..."
You: "No, Java the programming language"
Aria: "Java is an object-oriented programming language..."
Vaani learns from corrections within the conversation.
What Vaani Can Do#
✓ Speech Recognition
Understands multiple languages (32 total)
Works with accents and dialects
Handles background noise (with some limitation)
✓ Web Search
Current information (weather, news, facts)
Real-time data (stock prices, sports scores)
General knowledge questions
✓ Music
Find music on YouTube
Play audio through your speakers
Queue and playback control
✓ Conversation
Multi-turn dialogue
Context retention across messages
Personality-driven responses
Natural language understanding
✓ Customization
Language preference (32 languages)
Voice preference (male/female/other)
Response style (professional/casual)
API configuration
What Vaani Cannot Do#
✗ Real-time Actions
Vaani cannot:
Control your computer (open files, click buttons)
Send emails or messages
Make purchases or payments
Download files
Execute arbitrary commands
✗ Complex Analysis
Vaani cannot:
Analyze images or documents you show
Process large data sets
Perform complex calculations (though it tries!)
Debug your code (limited capability)
✗ Persistent Storage
Vaani cannot:
Remember conversations between restarts
Access your files or personal data
Access the internet beyond web search
Store files or create documents
✗ Advanced Functions
Vaani cannot:
Make phone calls
Send text messages
Control smart home devices
Access private accounts (email, banking)
(Some of these may be possible with custom development—see customization)
Troubleshooting Common Issues#
“Vaani doesn’t hear me”
Solutions:
Move closer to the microphone
Check microphone is enabled:
Settings→SoundReduce background noise
Check microphone volume in system settings
Verify microphone works:
python3 test_audio_direct.py
“Vaani misunderstands my accent”
Try:
Speak clearly and slowly
Rephrase after Vaani misunderstands
Change language setting if you’re bilingual
Check
docs/configuration.rstfor speech recognition options
“Music won’t play”
Check:
System volume isn’t muted
Audio device is working:
python3 test_music.pyYouTube is accessible in your region
No firewall blocking YouTube access
“Vaani gives wrong information”
Remember:
Web searches can be inaccurate
Vaani doesn’t verify information
You should verify important facts
Vaani might misunderstand your question
See Troubleshooting for more help.
Customization#
Change how Vaani behaves:
Change Language
# Interactive setup
python3 -c "from vaani_assistant.config import global_config; global_config.setup_initial()"
# Or edit .env file
echo "VOICE_LANGUAGE=es" >> .env
See Configuration for all options.
Change Voice
# Edit .env
echo "TTS_ENGINE_VOICE_ID=female_deep" >> .env
See Configuration for voice options by language.
Adjust Response Style
Configuration includes personality settings:
echo "ASSISTANT_PERSONALITY=formal" >> .env
# or
echo "ASSISTANT_PERSONALITY=casual" >> .env
See customization for detailed options.
Tips and Tricks#
Phrasing Questions
Instead of vague queries, be specific:
✓ "What was the first iPhone model?"
✗ "Tell me about iPhones"
✓ "What's the population of Tokyo?"
✗ "Tell me about Tokyo"
✓ "Play songs by David Bowie"
✗ "Play some rock music"
Waiting for Responses
Vaani needs time to:
Recognize your speech
Send query to AI engine
Optionally search the web
Generate response
Convert response to speech
Typical response time: 2-5 seconds depending on query complexity.
Handling Misunderstandings
When Vaani misunderstands:
✓ "No, I meant..." (gives context)
✓ "That's not what I asked" (corrects Vaani)
✓ Rephrase your question completely
✗ "You're wrong" (Vaani can't improve from this)
Music Quality
Music quality depends on:
YouTube source quality
Your internet bandwidth
Your audio device
Your speaker system
Expect MP3-quality audio (128-320 kbps).
Long Conversations
Vaani remembers conversation context, but:
Memory accumulates during a session
Context is reset when you restart Vaani
Very long conversations (100+ messages) may slow response
See Memory and Conversation Context for technical details
Advanced Configuration#
For power users, see:
Configuration - All configuration options
customization - Adapting Vaani for your needs
project_structure - Understanding the code
Voice and Audio System - Audio input/output details
Getting Help#
If something isn’t working:
Check Troubleshooting guide
Review Extended Frequently Asked Questions for common questions
Check logs:
cat logs/error.logEnable debug logging:
LOG_LEVEL=DEBUG python3 main.py
See Troubleshooting for detailed help.
Next Steps#
Explore customization to adapt Vaani
Read Voice and Audio System for audio details
Check How Vaani Thinks for how responses are generated
See Memory and Conversation Context for conversation handling