A voice-based assistant powered by OpenAI's GPT-4o-mini (openai), SpeechRecognition,playsoundand GTTS built in python.
the 'quirky' spelling of 'Tamborine' was intensional
-
Speech-to-text input using your microphone
-
GPT-4 conversation with memory
-
Continuous conversation loop
-
Voice commands
-
Handles errors gracfully
-
Microphone
-
Python dependencys in
requirements.txt:
openai
speech_recognition
gtts
dotenv
playsoundound-
Windows 10 or higher
-
Wi-Fi connection
- Open versions tab
- Locate v2.0.0
- Download and run
tambarine.exe
Clone the repository
git clone https://github.com/hippogriff101/Tambarine-AI.git
cd Tambarine-AI
Install dependencies
pip install -r requirements.txt
When running the python script locally you will need to provide your own open ai key via a .env!
You can get your key from https://platform.openai.com/account/api-keys If you need a free key head to this repo by @dan1471
Run:
python main.py
or
tambarine.exe
What you need to do:
- Speak clearly into your microphone
- The assistant will transcribe what you say and reply using GPT-4o-mini
- Say "stop" to end the session
This project obviouslly uses AI but I want to keep the use of AI in the develoment of projects transparent.
- Cleaning up mistakes in this
README.md - Polishing and debuging code
- Creating logo and bannar image
- Testing code (APi calls)
- Make a GUI version
- Option to use different ai models
All content is created using GPT-4o-mini Program exacutable has only been tested on windows Tambarine folder is created inin
Copyright (c) 2025 Freddie