DocTR

Use the DocTR OCR model through our Serverless Cloud API

DocTR is a document OCR model deployable via our Serverless Cloud API.

Code sample

Run DocTR through the HTTP endpoint directly with curl, or with the inference-sdk wrapper.

HTTP (curl)

1

Get your API Key

Create a Roboflow account, find your key on the Roboflow API settings page and make it available to your shell:

export ROBOFLOW_API_KEY="your-key-here"

2

Run the model

Call the /doctr/ocr endpoint with curl:

curl --location 'https://serverless.roboflow.com/doctr/ocr' \
  --header 'Content-Type: application/json' \
  --data '{
    "api_key": "'"$ROBOFLOW_API_KEY"'",
    "image": {"type": "url", "value": "https://media.roboflow.com/inference/license_plate_1.jpg"}
  }'

SDK (Python)

1

Get your API Key

Create a Roboflow account, find your key on the Roboflow API settings page and make it available to your shell:

export ROBOFLOW_API_KEY="your-key-here"

2

Install the dependencies

This package calls the model:

pip install -U inference-sdk supervision

3

Run the model

Run DocTR on an image containing text:

import os
import supervision as sv
from inference_sdk import InferenceHTTPClient

# Sample image containing text
image = sv.load_image_from_url("https://media.roboflow.com/inference/license_plate_1.jpg")

client = InferenceHTTPClient(
    api_url="https://serverless.roboflow.com",
    api_key=os.environ["ROBOFLOW_API_KEY"],
)

result = client.ocr_image(inference_input=image, model="doctr")

print(result["result"]) # Extracted text

The code above prints inference results to the terminal:

Mr
AUTPMATIC
280SE
34 T6511

Inference speed

Latency measured with Roboflow Inference on 1x NVIDIA L4, batch size 1, mean after warmup.

Model	Latency (ms)
doctr	83.5

Measured on a full document image (text detection plus recognition).

Set api_url to match your deployment target:

• https://serverless.roboflow.com for the Serverless Cloud API.

• http://localhost:9001 for a local Inference server.

• Your Dedicated Deployment URL for a private endpoint.

Use with Inference (self-hosted)

DocTR is a core model in Roboflow Inference, so it also runs on a server you host yourself. Start a local server, then point the same ocr_image call at it:

pip install inference-cli
inference server start  # serves http://localhost:9001

import os
from inference_sdk import InferenceHTTPClient

client = InferenceHTTPClient(
    api_url="http://localhost:9001",
    api_key=os.environ["ROBOFLOW_API_KEY"],
)

result = client.ocr_image(inference_input="./container.jpg")
print(result)

The response contains the recognized text and the inference time:

{'result': 'MSKU 0439215', 'time': 3.87}

Further reading

• How to detect text in images with OCR