Christopher

This patch uses the node.script-object to access the ChatGPT API. The responses are visualized using Jitter, rendering full control over the display. By accessing the Deepgram API for speech-to-text-conversion it's also voice activated.

This patch can be used create actual conversations with an AI, for example in a theatrical context. By framing communication on the patching level (for example by appending to messages sent to the AI model) one can also create more individual settings. In this demo, a simple line like "Answer as if you were Shakespeare", let's me talk to the resurrected bard on stage.

This patch uses the node.script-object to access the ChatGPT API. The responses are visualized using Jitter, rendering full control over the display. By accessing the Deepgram API for speech-to-text-conversion it's also voice activated.
This patch can be used create actual conversations with an AI, for example in a theatrical context. By framing communication on the patching level (for example by appending to messages sent to the AI model) one can also create more individual settings. In this demo, a simple line like "Answer as if you were Shakespeare", let's me talk to the resurrected bard on stage.

This patch uses the node.script-object to access the ChatGPT API. 

shakespeare-talkbot-with-chatgpt-and-jitter

Thank you for the great video! Did you ever come across the issue of the max2gpt.js node script outputting random responses even though there was no prompt sent to the API? Either through a manually inserted message (prompt) or through the deepgram.js node script?

Shakespeare Talkbot with ChatGPT and Jitter

2023

Mannheim