About The Connector & Demo
This Forge component comprises a Connector for Google Gemini Pro Vision. We've also included a demo application to demonstrate how to utilize the connector, enabling you to execute prompts on a given image and receive responses. The demo application is straightforward, and we encourage you to explore your own use cases for addressing business challenges. The objective of the connector is to be easily integrated into your current development workflow.
Pre-requisite
Here is the step-by-step documentation for getting the API key for Gemini Pro Vision model.
Click on the below URL to proceed further
Build with the Gemini API | Google AI for Developers
Then click on Get API key in Google AI Studio. And try to login with your Gmail account. If you don’t have then create a new one.
Then Click on Get API key in left corner to get your free API key.
Outsystems configuration for the connector
Outsystems do not support Gemini Pro Vision’s generated JSON, however we can overcome by adding JSON in the Site Property
The below JSON is already added in Local Variable's default value of the connector application
{
"contents":[
"parts":[
{"text": "Question"},
"inline_data": {
"mime_type":"model",
"data": "base"
}
] } ] }
Configuring OutSystems Demo Application
Then Add that key as Site property of our demo application to continue our services
About the Demo Application
In the demo screen,
You have to select an image from your local drive to upload.
Upload that image using the icon in the demo application.
Select the image(in case of multiple files) on which you need to ask questions.
You can you prompts and the answer will be shown from that document itself.
Note: Under the cover, the Gemini Pro Vision Generative AI Model will fetch you the answer.
Then add below JSON in Site Property of the connector application using service center
Because service studio won’t accept JSON as input to site property so add that value via service center