The public API is now available for all users who have purchased a V-PRO subscription!

To use the API, generate a key in your personal account page. Use the API key when calling API methods through the Authorization header:

Authorization: Bearer YOUR-API-KEY

The main API method is the recognizeFile file recognition method. It takes as a required parameter the file that needs to be recognized as input and returns information about:

  • recognition identifier. Used to refine the list of keywords
  • base64 encoded file preview
  • information about the remaining limit
  • recognized keywords

Each successful method call reduces the number of remaining recognitions by one.

To send a video for recognition, first split it into separate frames (up to 100 frames) and send them all at once to the recognizeFile method. In this case, the call will also be counted as one recognition

Files sent for recognition will be processed one by one. The minimum delay between requests is 20 seconds. If you want to process files more often, write to us via the feedback form

The second important method is updateKeywords. It is used to get more accurate keywords. You can pass it a recognition identifier (valid up to 12 hours) and arrays of positive (i.e. known to match) and negative words (not matched to the image). Calling this method does not consume the recognition limit

For generating of descripton use describeFile method. It receives recognition id and file preview in base64 format, which you can get from recognizeFile method. If you have passed mutliple files into recognizeFile method, it will return not only preview, but also describingPreview field, which you need to use for description generation

Available methods are described in the documentation. Feel free to ask questions about the API via the feedback form

API usage examples: