Whisper-Transcribe-and-Summarise

A webserver that transcribes and summarises audio using ollama llama3 and outputs to notion and discord

Audio Transcription Web Server

This project is a web server that transcribes audio using OpenAI's Whisper model, generates a summary and a title with Ollama's LLaMA3 model, and outputs the results to both Discord and Notion.

Features

• Audio Transcription: Transcribes uploaded audio files using OpenAI's Whisper.
• Text Summarization: Generates summaries and titles for the transcribed text using Ollama's LLaMA3 model.
• Discord Integration: Sends transcribed and summarized text to a specified Discord webhook.
• Notion Integration: Creates a new page in Notion with the transcribed text and summary.

Installation

1. Clone the repository.

2. Install the required dependencies:

   pip install flask
   pip install requests
   pip install json
   pip install os
   pip install subprocess
   pip install argparse
   pip install whisper
   pip install ollama

Configuration

1. Discord Webhook: Set the Discord webhook URL in the post_to_discord function.
2. Notion API: Obtain a Notion API key and set the NOTION_KEY variable.
3. Notion Parent Page ID: Find the Notion parent page ID and update the page_id variable in create_notion_page.

Usage

1. Start the web server:

   python server.py

2. Access the server's upload page at http://localhost:1987/.

3. Upload an audio file to start the transcription process.

Contribution

Contributions are welcome! If you find a bug or have suggestions for features, please open an issue or submit a pull request.

License

This project is licensed under the MIT License.