This is a FastAPI-based web application that:
Accepts user-uploaded images
Takes natural-language instructions
Automatically generates a structured prompt from the instructions
Analyzes the image based on the generated prompt
Returns whether the image is compliant or non-compliant based on user specifications
Logs all responses and token usage
- Upload an image and describe what you want analyzed
- Automatically generates a structured prompt based on your input
- Performs compliance or classification analysis
- Logs image name, response, and token counts
- Built with Docker & Docker Compose
- Set up your environment variables
Create a .env file in the root of the project:
GOOGLE_APPLICATION_CREDENTIALS=google_credentials.json
PROJECT_ID=your-google-cloud-project-id
MODEL_NAME=your-model-name