SpotifyTranscripts - Open Source AI-Powered Podcast Transcriptions and Chapter Segmentation

Spotify Transcripts: Bringing Podcasts to Life with AI

✨ Key Features

Spotify Transcripts offers a unique experience for podcast enthusiasts by leveraging artificial intelligence to enhance podcast accessibility and organization. Here are the core features:

Transcripts: Converts spoken words in podcasts to text with precise timestamps using advanced speech recognition technology.
Search: Enables users to search through the transcripts to find and jump to specific parts of a conversation quickly.
Chapters: Automatically divides podcast episodes into chapters centered around key topics discussed, providing structured navigation.
Subtitles: Boosts accessibility for those who are hearing impaired by providing readable subtitles of podcast content.

📖 About the Project

Spotify Transcripts is a culmination of previous projects aimed at enhancing the usability of podcasts. Initially, the developer participated in Spotify’s summer hackathon, creating a tool for topic-specific navigation within podcasts. Later, they focused on generating subtitles for podcasts, a feature heavily requested by the community. Inspired by the advancements in AI technologies in 2023, the developer amalgamated these projects to produce a superior podcast player using Open AI’s APIs.

Spotify itself later introduced a similar feature set, validating the project's innovative approach.

⚙️ Technologies Used

To develop Spotify Transcripts, several advanced technologies were employed:

React: Utilized for building the attractive and responsive front-end interface.
Tailwind: A CSS library providing stylish and customizable designs.
Python: Powers the backend processes, especially for handling transcription logic.
Flask: Connects the frontend and backend, facilitating smooth data transfer and processing.
Spotify & Google Speech Recognition APIs: These APIs gather podcast information and convert speech to text.
Open AI's GPT 3.5 API: Uses AI to segment and organize transcripts into chapters.

The developer used this project as a learning opportunity to bridge React frontends with Python backends, overengineering a personalized API to manage transcription processes.

🚫 Limitations

Despite its innovative features, Spotify Transcripts faces limitations due to Spotify’s API restrictions, which currently only allow access to 30-second podcast previews. As such, this project stands as a proof-of-concept.

🚀 Getting Started

To experiment with Spotify Transcripts, follow these steps:

Sign Up for API Keys: Obtain developer API keys from Spotify and Open AI.

Configure API Keys: Add your keys to a .env file in the root directory of the project:

REACT_APP_SPOTIFY_CLIENT_ID=YOUR_SPOTIFY_CLIENT_ID_GOES_HERE
REACT_APP_OPEN_AI_KEY=YOUR_OPEN_AI_KEY_GOES_HERE

Run the Project: Initiate the frontend and backend components in separate terminals:

Backend:
```
export FLASK_APP=backend
export FLASK_DEBUG=1
flask run
```
Frontend:
```
cd frontend
npm start
```

🎞️ Demo and Screenshots

For a visual demonstration, watch a 1 min demo. The project includes an attractive user interface with various screens, such as:

Home and Discovery Pages with Spotify authentication.
Episode screen showcasing subtitles.
Detailed chapters overview and a search feature for transcripts.

By integrating these technologies and features, Spotify Transcripts showcases how AI can transform how users interact with podcasts, making them more accessible and organized than ever before.