Recognition Request

A recognition request has the following fields:

API fieldData typeDescriptionExample valuesRequired
device_idstringThe ID of the device making the recognition request.DEVICE1, 123456Mandatory
imagesarray of objectsAn array of objects.n/aMandatory
camera_typestringThe type of camera the image was captured on. Must match one of the camera types the device has been registered with.generic, MP7000-P, MP7000-L, 7879e-C3, TC1, TD0, TDI, SM6000Mandatory
imagestringBase64 encoded .png or .jpeg image. For more information on image format, see Image format.Too long to provide exampleMandatory
capture_timedate-timeTime the image was captured, in ISO format.2024-01-04T12:30:45Z, 2025-07-15T12:30:00Z, 2028-12-01T23:55:00Z, 2030-03-14T15:20:15ZMandatory
weight_gramsnumberWeight of the product in grams. Used to improve results.150, 100.1Optional
include_scoresbooleanWhether to include confidence scores in the response. Recommended for experienced users only.true, falseOptional
score_thresholdnumberCustom confidence score threshold for the response. Can result in unexpected behaviour. Recommended for experienced users only.0.9, 0.8, 0.88Optional
max_optionsintegerMaximum options that can be displayed to the user. Default is 4 if omitted.1, 2, 3, 4, 5, 6Optional

Here's how to make a recognition request using a Unix-like command-line interface:

curl --request POST --url <> --header 'accept: application/json' --header 'content-type: application/json' --header 'tiliter-api-key: YOUR_API_KEY' --data ' { "images": [ { "camera_type": "TDI", "image": "BASE64_IMAGE", "capture_time": "2024-01-04T12:30:45Z" } ], "device_id": "DEVICE1"} '

The expected response to this is:

{ "id": "57355cd5-072c-45f3-84ba-255306ee961c", "product_recognition": { "result_type": "recognised", "options": [ { "product_id": "PRODUCT1", "quantity": null, "score": null } ] }, "bag_recognition": { "present": false } }