BassGPT

• 

  Billiam

• 

  Key On

• 

  Key Off

• 

  Another pic of the lad

• 

  Inside the buttons

• 

  Billy Brains

• 

  Billy guts

Inspiration

We came across Billy in a time of need. Decided to help him reel in his full potential and become Billiam Bass the BassGPT. Now he's ready to "dive deep" and make a splash in the tech world, "hooking" everyone with his fishy wisdom!

What it does

With the turn of a key and a press of a button, you're able to ask BassGPT anything that your heart desires. "Billiam" will enthusiastically respond accurately with a quirky personality, providing companionship from the deep blue sea.

How we built it

BassGPT contains 3 systems:

1. Button Activated Switch System serially records user input with an Arduino to BassGPT. By turning a key switch, an Arduino serially signals BassGPT to popup a GUI for the user to input their questions. Pressing the button triggers the program flow - making the process as smooth as a swim in the deep blue.
2. frankenfish, controlled by an ESP8266 handling motor control, is hidden inside the fish mount. By running a web server, it waits for and interprets motor commands sent by BassGPT via HTTP requests to hijack the built-in motors to reanimate poor Billiam's body for our amusement...
3. BassGPT brings it together in Java. It communicates with Gemini's API for the user's conversation, then pipes its response directly into ElevenLab's text-to-speech API. Then with some audio processing, we generate motor commands using the audio output to emulate a talking tuna!

Challenges we ran into

Our biggest challenge was figuring out how to work the motors without a driver or shield. With a lot of trial and error and with the big help of MLH desk we were able to leverage transistors and other spare parts to get everything up and swimming.

Accomplishments that we're proud of

We were the proudest the first time we were able to make this fish's motors to move

What we learned

A lot of hardware, Arduino's, circuits, communicating between hardware and software, API's, audio processing, and having other programs be able to interpret serial signals from the Arduino IDE code.

What's next for BassGPT

If we had more time, utilize on board audio processing to limit the use of API's, we would implement a implement a microphone so that you can directly talk with the fish, we would put a new speaker inside of the mount rather than relying on external Bluetooth speaker, and we would add motor shields and drivers so we could have better control and power distribution so that we could ideally also use the tail.

Built With

• api
• arduino
• c++
• esp8266
• gemini
• hardware
• java
• webserver