Voiceai is a library for controlling a machine using text and voice based interface. It uses

1. The Stanford NER tagger for NER Tagging
2. The Stanford POS tagger for POS Tagging
3. The Facebook fastText for text classification
4. The pint library for converting units
5. The fixer.io api for converting currencies
6. The duckduckgo wrapper for searching

1. Music control

• Play Song/Artist/Album
• Pause
• Stop
• Resume

Music library must be provided as json file.

Note : ML Algorithm for learning taste of music coming soon.

2. Hardware Control

• Adjust volume
• Adjust brightness (needs root access)

3. Conversion Control

• Convert units and dimensions
• Convert currencies (internet required)

4. Web search Control

• Search anything using the DuckDuckGo api

5. Greeting Control

• Engage in a casual conversation with the bot

6. Alarm Control

• Set alarms and reminders

7. Wolfram Control

• Use Wolfram Alpha to get answers to mathematical answers

1. Training

• Add new sentences for text classification (fastText)
• Add new NER tags for better support (Stanford-NER)
• Add new POS tags for better support (Stanford-POSTagger)

Place the voiceai directory in your project

git clone https://github.com/vidursatija/voiceai.git

import voiceai
VC = voiceai.VoiceAIControl()
print(VC.process_message("Play some Taylor Swift songs"))
print(VC.process_message("Increase brightness by 10%"))
print(VC.process_message("How many miles are there in a kilometer?"))