New Google Gemini AI Piece

Activepieces + Google Gemini: Supercharge your flows with AI!

Hello Activepiecers, we are excited to announce a new piece: Google Gemini.

:memo: Text & Code Generation
Create various content types like emails, blog posts, scripts, and poems.
:globe_with_meridians: Language Translation
Instantly break language barriers and connect with diverse audiences.
:speech_balloon: Personalized Interactions
Craft unique responses, emails, and chatbots for authentic customer engagement.
:rocket: Content Creation Speed
Generate outlines, ideas, and descriptions swiftly, saving time for strategic planning.

List of actions for this piece:

  • Generate Content: You can pass the model a prompt to generate content from. This uses the gemini-pro model.

  • Generate Content from Image: You can send an image and a prompt for the model to interpret them. The image can be an URL of a base64 representation (with the mime type included). This uses the gemini-pro-vision model.

  • Chat Gemini: You can have a conversation with the model passing a “Memory Key”. AP will use that memory key to store a JSON that will be passed to the model every new message. If you want to start a new conversation just pass a new Memory Key. If you want the model to no remember anything, just don’t pass any Memory Key (keep the field blank, this is equivalent to the “Generate Content” action). This action is limited to the gemini-pro model. According to Google:

The gemini-pro-vision model (for text-and-image input) is not yet optimized for multi-turn conversations. Make sure to use gemini-pro and text-only input for chat use cases

I’ll keep myself posted on this and add chat capabilities using images when Google tell us is OK to do that.

Remember you can get your API key from:


This is awesome! @PFernandez98 Any idea how to get access to Gemini?

I don’t remember if I got some kind of early access (I don’t think so tho), but you can get your API Key in here: and the docs are in here: Gemini Docs and API Reference  |  Google AI for Developers

I think I got it, I had to turn on access to Early Access apps in our Google Workspace account. I’ll try it out soon!

1 Like