We’ve got ourselves a whale of a Raspberry Pi project on our hands today. A maker and developer known as TheAmateurRunner over on Reddit has brought the classic Big Mouth Billy Bass into the modern era, with a little help from our favorite SBC. If you’re not familiar with this talking fish, it was a popular wall-mounted decoration from the early 2000s, known for singing and generally startling anyone who walked by, until the motion sensor was swapped out for a button.
According to TheAmateurRunner, this is very much still a work in progress. But so far, they’ve managed to control the Big Mouth Bill Bass with a Raspberry Pi and successfully integrated ChatGPT. It flips its tail and turns its head to relay messages retrieved from ChatGPT. At the moment, the mouth doesn’t move, but this should be fixed with future updates.
TheAmateurRunner explains that the Raspberry Pi communicates to a couple of APIs to handle the interactive features. It uses both Google and OpenAI to receive voice commands as well as process responses for replies. The idea is to modify a Big Mouth Billy Bass that can talk back and forth with users using ChatGPT.
If you don’t want to disassemble a working Big Mouth Billy Bass, we don’t blame you. In this case, TheAmateurRunner is using a broken one purchased years ago and restoring it with new functionality. The Raspberry Pi used in this project is a 3B+ but you could get away with using a Raspberry Pi 4 if that’s what you have on hand.
To receive commands, the Pi is using Google Speech to Text. This text is then parsed to OpenAI for processing. Responses are handled using Google Text to Speech. TheAmateurRunner explains in a comment that the unofficial ChatGPT API wasn’t working, so a workaround had to be introduced with OpenAI.
If you want to get a closer look at this Raspberry Pi project in action, check out the thread shared to Reddit. There, you can also get a glimpse of the hardware setup controlling the Big Mouth Billy Bass. We’re looking forward to future updates on this project so be sure to follow TheAmateurRunner to see the final project.