Image input
Send an Uzbekistan Texpasport image
Source imageDetected fields
AI Computer Vision + OCR
Read a Texpasport
Read the plate number, VIN and owner details from an Uzbekistan vehicle registration certificate image. Verse uses AI Computer Vision and OCR to turn the text into structured data for your software.
texpasport.jpg · ExampleCompleted
O'ZBEKISTON01
TEXNIK PASPORT
01 A 123 BC
Data by field
plate_number99%
01 A 123 BC
vin98%
XW8ZZZ61ZJG...
owner96%
M. Tursunov
schema_version: texpasport-v1
Computer Vision + OCR+JPEG / PNG+Structured JSON+Prepaid balance: $0.003
From image to text
Currently supports one Uzbekistan Texpasport format
Field detection
Computer Vision finds vehicle and owner details in the image. The result schema has 20 fields, including plate number, VIN, model and year.
Text recognition
OCR turns text in the image into field values. JSON organizes them for your software; check the results against the image.
"plate_number": {
"value": "01 A 123 BC",
"confidence": 0.99,
"bbox": [40, 80, 220, 120]
}01
Send
Send a JPEG or PNG image up to 5 MB in a POST request. Use an API key for access and an idempotency key to avoid creating duplicate requests when retrying.
02
Read
Verse processes the image in the background: Computer Vision locates fields and OCR reads their text. The result is not returned immediately.
03
Retrieve
Check the request status with repeated GET requests until processing finishes. Retrieve the JSON result and review the extracted values against the image.
Less manual entry
Use extracted Texpasport data to help fill forms and records. Staff should review the results before using them.
Government
→
Prefill vehicle details in service application forms for staff to review.
Insurance
→
Copy vehicle details into insurance applications and claim forms for review.
Logistics
→
Add vehicle details to fleet records without typing each field from scratch.
Fintech
→
Prefill vehicle details in loan applications. OCR does not verify identity or document authenticity.
Example result
Data in JSON
This is a shortened illustrative response with fictional data, not the full API response. JSON groups the extracted fields for software to read. Confidence is a model score, not an accuracy guarantee; human review is needed.
View docs ↗GET /api/v1/extraction-requests/{request_id}
{
"schema_version": "texpasport-v1",
"fields": {
"plate_number": {
"value": "01 A 123 BC",
"confidence": 0.99,
"bbox": [40, 80, 220, 120]
},
"vin": { "value": null, "confidence": null, "bbox": null }
}
}Common questions
Which documents are supported?+
Currently, Verse supports one Uzbekistan vehicle registration certificate format, commonly called a Texpasport. It does not offer universal support for other formats, countries, or document types.
Which fields are in the schema?+
The schema has 20 fields: plate_number (registration plate), model (vehicle model), color, owner, address, issued_date (issue date), issuing_authority, pin (personal identification number), year (year of manufacture), type (vehicle type), vin (vehicle identification number), maxmass (maximum mass), mass (vehicle mass), engine_num (engine number), engine_cap (engine capacity), fuel (fuel type), seats (seating capacity), stand_place (standing capacity), special (special notes), and series (document series). Not every image yields every field; missing values are null.
Which files can I send to the API?+
Send a JPEG or PNG image of the supported Uzbekistan Texpasport format, up to 5 MB per file.
What does the extraction API return?+
The JSON result contains value (the text read), confidence (the model's confidence score), and bbox (coordinates of the box around the field in the image). Missing values are null, meaning no value was found. Confidence is not a guarantee that the text is correct.
How do I connect the API?+
Create a request with POST, using your API key for access and an idempotency key to avoid duplicate requests on retries. Processing runs in the background. Use repeated GET requests to check the status until processing finishes, then retrieve the JSON result. This repeated checking is called polling.
Do I need to check the results?+
Yes. Compare the extracted data with the source image before using it. Text may be missing or misread, even with high confidence. OCR does not establish document authenticity, verify identity, or replace KYC (customer identity checks).
How much does Texpasport OCR cost?+
Payment uses a prepaid credit balance. Each completed image costs $0.003, including empty or low-confidence results. If processing fails, the reserved credit is released back to your available balance.
Pricing
$0.003per image
Top up your prepaid balance. Each completed image costs $0.003, even if the result is empty or has low confidence. If processing fails, the reserved credit is released back to your available balance.