🤖 AI Terminal Assistant

Conversational AI assistant powered by the Groq API. Features streaming responses and conversation memory.

Setup

pip install groq colorama SpeechRecognition pyttsx3 pyaudio

Note: pyaudio needs a C compiler on Windows (or Microsoft C++ Build Tools) to build from source if no prebuilt wheel exists for your Python version. Voice input (/voice) requires it; everything else, including voice output (/speak), works without it.

Set your API key (get one free at console.groq.com):

setx GROQ_API_KEY your_key_here     # Windows (permanent)
$env:GROQ_API_KEY = "your_key_here" # Windows (current session)

Run:

python ai_assistant.py

Commands

/clear            new conversation
/save [name]      save conversation to conversations/<name>.json
/load [name]       load a saved conversation
/role [name]       switch system role: assistant | coder | translator | poet
/voice             record a message from your microphone and send it
/speak             toggle text-to-speech for AI responses
/help              show help
/exit              quit

Features

• Streaming responses (token by token)
• Conversation memory (last 10 messages)
• Save/load conversations as JSON
• Colored terminal output (colorama)
• Token count and cost estimate per message
• Voice input via microphone (speech_recognition, Slovak language model)
• Voice output / text-to-speech (pyttsx3)

Built with

Python 3 • Groq API • llama-3.3-70b-versatile