Getting Started

Welcome to our API.

Convert images to feature vectors for use in downstream applications.

This API is still under development and will evolve.

What can the vectors be used for?

  • Finding similar images
  • Ranking for recommender systems
  • Clustering images to different categories
  • Classification and other machine learning tasks
Logo

Errors

Code Name Description
200 OK Success
500 Error Malformed request or invalid base64

All errors will return JSON in the following format:

{
  "statusCode": 500,
  "message": "Internal server error"
}

Scaling

Need something faster and more scalable than our free tier?

Check out the paid tier

/image

Get a feature vector for any image

Parameters
image
The base64 encoded image

Response time varies in the free tier - normally 1-3 seconds.

from base64 import b64encode
import json
import requests

url = "https://sruhyjfrd6.execute-api.us-east-2.amazonaws.com/image"

with open("your_image_file.jpeg", "rb") as imageFile:
    img_string = b64encode(imageFile.read()).decode('utf-8')

data = {"image": img_string}
r = requests.post(url, json=data)

if r.status_code == 200:
    # Get latent vector
    latent_vector = r.json()['body']['image_vector']
    print(len(latent_vector))
else:
    # Print error response
    print(r.json())
{
"statusCode": 200,
"body": {
  "image_vector": [
    -0.01137700118124485,
    -0.0895775631070137,
    ...]
  }
}
{
  "statusCode": 400
  "image": null,
  "error": "Invalid base64 string"
}