There is now an alternate flow available for interacting with Googles Gemini-Pro model that supports multi turn conversations that can be found here
Requires Node.js v18+
Requires installation of @google/generative-ai npm module
to install it run
npm install @google/generative-ai from the directory where Node-RED is installed.
Requires an API KEY that can be obtained from here if you live in a supported country.
The rate limit for Gemini Pro models is 60 requests per minute (RPM), which should be plenty for most use cases.
You can adjust the maxOutputTokens in the Gemini function node to fine tune responses.