This code provides an example of how to create an end-to-end solution to inference images of food products, detecting the name and brand. Additionally, it allows you to create a recipe from text, including images. It has been optimized to read images that contain a single product only
The infrastructure is based on:
- Amazon API Gateway - Entry point to send images and text
- Amazon Lambda Function - Handle requests to detect images and generate recipes
- Amazon Cognito - Provides authentication for the API
- Amazon Bedrock - Used for inference requests to Anthropic Claude 3 sonnet and Amazon Titan image models
- Amazon S3 - Host the web application
- Amazon Cloudfront - Content delivery network for the web application
Most of the deployment of the infrastructure is automated using the Serverless Application Model. For detail instructions, follows the steps here
Example single page application built in Vue JS - it is integrated with Cognito to provide a seamlessly experience, allowing users to sign up and sign in.
Detail instructions to install the web app can be found here
Please open an issue for any questions or issues