This is a fun app that let you turn your thoughts into tweets in a jiffy!

You simply talk to the mic and it records your voice, transcribes it and turn it into a tweet with a click of a button. You can run the ASR Whisper models either locally or via OpenAI API. By default it's using the API, but if you'd like to use the models locally, check audio_service_local.py file. There are two different approaches to run the model locally using two different python libraries: fast-whisper and insanely-fast-whisper. Pick whichever that suits you or faster on your system, and update the FastAPI endpoint/path accordingly.

1. Clone the repo:

git clone https://github.com/mallahyari/transonic.git

2. To run the frontend, go to the frontend directory and run:

npm install
npm start

You can skip the authentication if you run it locally. Nevertheless, if you would like to set up authentication, you'll need to add your CLERK_PUBLISHABLE_KEY to your .env file in frontend folder.

Additionally, if you tend to run FastAPI backend locally, make sure to update the BACKEND_URL inside the frontend/src/config.ts to the appropriate url, e.g. http://localhost:8000.

3. To run the backend, from backend/app folder run:

pip install -r requirements.txt
python main.py

If you plan to use OpenAI Whisper API, you will need to enter your OPENAI_API_KEY. To do that, simply create a .env file in the backend/app directory with the following info:

OPENAI_API_KEY=

• App logic: React + ant design
• Authentication: Clerk
• Deployment: Vercel

• App logic: FastAPI
• Deployment: Fly.io

If you have any comment or feedback, please don't hesitate to reach out directly or use the Discussions section and open a new topic. You can also reach out directly via Linkedin or Twitter.