Tars: Voice AI Assistant 🚧

TARS is an early-stage Voice AI Assistant built to process Speech-to-Text (STT), Text inputs and generate both Speech-Based and *Text-based responses. It integrates features such as file handling, news retrieval, minimal web search, and persistent conversation tracking.

---

NOTE: Tars is built as a learning project

---

Features

• Speech-to-Text (STT) Processing for voice input
• Text-to-Text Can chat with the model directly
• Text-to-Speech (TTS) based responses.
• Persistent conversation storage using SQLite
• Integrated Tools:
  1. Websearch - Uses DuckDuckGo library. This tool can search for Pages, Images, Videos and news (Try to set the max_results to low to reduce high token consumption)
  2. Wikipedia - Can do wikipedia Search, Summary and Content(Using tool will consume a lot of tokens.)
  3. News - Fetches news based on country and categories
  4. File handling - Can read, write, search and open files (For example You can ask it open video.mp4 in downloads folder and it can open the file.)
  5. Video downloading utilities
• Conversation history tracking
• Automatic conversation summarization

---

Installation

Prerequisites

• Python 3.12.x -- Tested in this environment

Clone the repo

git clone https://github.com/Praneeth-Gandodi/Tars.git

Navigate to project folder

cd Tars

Create a Python virtual environment

python -m venv .venv

Activate virtual environment

If Windows:

.venv/Scripts/activate

If Linux/Mac:

source .venv/bin/activate

Install dependencies

pip install -r requirements.txt

---

Configuration

This project requires a Groq API key, which you can obtain for free from the official Groq website.

• You will be prompted to paste your Groq API key and model name.
• Enter your Groq API key and the desired model name when prompted.
• This setup is required only once. If you want to change the model name or API key later, open the generated .env file and update the values there.

---

Usage

Run the tars.py file:

python tars.py

---

Development Status : Paused⏸️

---

Implemented / Planned Features

• Text-to-Speech (TTS)
• Web search
• News
• Wikipedia
• File handling (Read and write text files, Search for files with names and types)
• Video Downloader(supports Youtube(Video/Audio), Instagram(Video), Facebook(Video))
• Weather
• Date & Time
• Speedtest
• Browser Control
• Spotify
• Application control
• Integrate Camu (for university attendance tracking; college-specific feature; can be ignored)
• Change the framework
• Automation of tasks

License

MIT License © 2025 Gandodi Praneeth Kumar