Guppy communicates with OpenAI’s API and performs chat completions using different GPT models that can be configured.
Conversations can be saved and loaded for documentation and continuation. Only JSON stores both the conversation and settings.
Design focus has been on simplicity and colorfulness, aiming more for the general population than specifically software developers.
To avoid cost, the user needs to enter their own OpenAI API Key, which admittedly is somewhat of a hurdle from an OpenAI registration point of view.
Guppy runs locally in the user’s browser, and communicates externally only with OpenAI and our own Matomo installation. No custom server-side code is used.
The core functionality was developed during a few days at the end of August 2023. Enhancements are still made based on need.