The Idea
There is a lot of information that is streamed across the airwaves, and it would be impossible for one person to listen to one or more radio station 24/7 to learn everything that is happening around them. Fortunately, we have technology to help us out!
With this project, we will capture radio broadcasts, transcribe them into text, summarize the content using locally hosted large language models (LLMs), and then publish them to a locally hosted website for consumption and further analysis.
High-Level Overview
- Record radio broadcasts 24×7 on a low-powered computer.
- Using a computer equipped with a GPU, transcribe the radio broadcasts using Open AI Whisper.
- Summarize the transcripts using a Large Language Model (LLM).
- Add raw transcripts and summaries to an Open Web UI knowledge base.
- Publish summary, full raw transcript, and playable audio file to a locally hosted WordPress instance.
Pre-requisite Tutorials
- How to Schedule Audio Capture via Line-in on Linux
- How to Set Up Passwordless SSH & SFTP Access Between Two Linux Servers
- How to Set Up Ollama and Open WebUI on Linux
- How to Set Up OpenAI Whisper for Locally Hosted Audio Transcription on Linux
- How to Set Up a Local LAMP Stack with WordPress on Debian-Based Linux