Ask Nova

🦾 ChatGPT+Whisper+ElevenLabs

A Typescript port (and extension) of the python web application "Ask Hugh" created by IgnoranceAI available here: https://github.com/IgnoranceAI/hugh

This is intended as a simple starting project for ChatGPT+Whisper+ElevenLabs.

It allows you to record audio, transcribe it, and then ask a question based on the transcription. The application uses the Whisper API, the ChatGPT API, and the ElevenLabs API.

📦 Install

1. Clone this repository: git clone https://github.com/karl-friman/asknova-ts.git

2. Install npm i

3. Duplicate the file .env.sample to a file called .env.local and add your API keys:
   OPENAI_API_KEY = "YOUR API KEY HERE".
   ELEVENLABS_API_KEY = "YOUR API KEY HERE"

4. Create a folder called "uploads" in the root directory.

✨ Usage

1. Start the server: npm start
2. Open your web browser and navigate to http://localhost:3000.
3. Click the “Record” button to start recording audio. Speak into your microphone and then click the button again when you’re done. The transcription will appear in the text box.
4. Alternatively, type your question into the text box and click the “Ask” button to submit your question and generate a response.
5. The response will appear in the area below the audio player. It will begin playing and typing once it is ready.

🙏 Credits

The original "Record and Ask" was created by Artificial Ignorance using Flask, OpenAI’s Whisper, OpenAI’s ChatGPT, and ElevenLabs.

🪪 License

This project is licensed under the MIT License - see the LICENSE file for details.