Project Objective
To optimize my personal education and efficiently manage time, this project focuses on automating the summarization of YouTube videos from selected channels. Utilizing the capabilities of Node-RED, it integrates YouTube for precise video selection, employs OpenAI's Whisper for accurate audio transcription, and leverages Notion for seamless organization. The core of this initiative is to transform extensive video content into concise, easily digestible summaries, optionally translated into German, ensuring that key educational insights are captured succinctly. This methodical process not only streamlines content absorption but also significantly reduces the necessity to watch each video entirely. The automation extends to updating Notion with these summaries, alongside pertinent video details, and proactively sending an email notification with the link to the newly updated Notion page. This comprehensive approach aims to enhance learning efficiency, enabling quick assimilation of crucial information from my preferred YouTube channels with minimal direct engagement.
Seeking Expertise
I am on the hunt for a seasoned Node.js developer to bring this vision to life on an AWS Lightsail-hosted Node-RED instance.
Functional Workflow
Periodic YouTube Monitoring: Set up a Node-RED trigger to check for new videos on designated channels hourly, utilizing the YouTube Data API. The setup should allow for easy duplication to monitor multiple channels.
MP3 Extraction: On identifying a new video, employ node-red-contrib-ytdl (or a similar module) to download the video's audio track in MP3 format.
Transcription via Whisper AI: Use an HTTP request node configured with the necessary authorization headers to submit the MP3 file to OpenAI's Whisper for transcription.
Language Detection and Translation: Assess the transcription's language. If not in German, translate using the Deepl API through node-red-contrib-deepl or an equivalent HTTP request node.
Content Summarization: Formulate a JSON object with the transcription (and its German translation) and utilize OpenAI's GPT (text-davinci model) for summarization.
Notion Database Update: With the summary and other pertinent information (e.g., video title, URL, summary date/time), update a Notion database, potentially requiring direct HTTP requests to the Notion API for authentication and data management.
Email: Send an email with the title of the YouTube video in the subject line and the URL to the new notion page in the body.
Cleanup: Post-process, remove the MP3 file from the server to maintain cleanliness.
Additional Considerations
API Keys Management: Instructions or best practices for securely managing API keys and tokens within Node-RED.
Error Handling Strategies: Guidelines on handling API rate limits, transcription inaccuracies, or any translation errors.
Performance Optimization: Suggestions on managing large mp3 files from long videos (up to five hours) efficiently without overloading the server.
Security Measures: Recommendations on securing the flow, especially when dealing with personal educational data and API interactions.
Inspirational Resources:
For a practical understanding, explore similar implementations here: and the GitHub repository here.
Please contact me to discuss your fee, eta etc.
Thank you very much for considering to work with me on this project.