Voice-Controlled Spot with Whisper and GPT

by Bart Trzynadlowski - March 23, 2024

Click here for video of final result.

Won the Robo-Hackathon at CircuitLaunch in Oakland today with this project. Unfortunately didn't have enough time to add proper streaming transcription support. Voice samples are sent in 5 second chunks, which means if a command straddles a boundary, it will likely be missed.

Thanks a lot to MerkleBot and DuploCloud for organizing this.