> For the complete documentation index, see [llms.txt](https://docs.roboflow.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.roboflow.com/models/supported-tasks/ocr.md).

# OCR

Read serial numbers, labels, and documents from images with OCR models that run pretrained on Roboflow.

## What is OCR?

OCR (optical character recognition) reads text in an image. The model returns the text it finds, and some models also return where each piece of text is. Accuracy is best when the input is a tight crop of the text, so a common pattern is to find the text region with an [Object Detection](https://docs.roboflow.com/models/supported-tasks/object-detection.md) model first, crop it, and then run OCR.

<figure><img src="https://3448603185-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fcmo9mhfIjYrvFFy1U7dk%2Fuploads%2Fgit-blob-467ca7e92f4c15e61d91511a8866cf44235cdc7e%2Fimg-task-ocr-wide.webp?alt=media" alt="A chalkboard price sign among clementines, with each line of text boxed and read out as Clementinen, 10 St., 4 €"><figcaption></figcaption></figure>

## Supported OCR models

Roboflow supports these OCR models. See the [OCR models](https://docs.roboflow.com/models/supported-models.md#ocr) table for export and deployment support per model.

<table data-search="false"><thead><tr><th>Model</th><th>License</th><th>Latest release</th></tr></thead><tbody><tr><td><a href="https://docs.roboflow.com/models/supported-models/glm-ocr.md">GLM-OCR</a></td><td><a href="https://docs.roboflow.com/models/model-licenses.md#other-models">Upstream license</a></td><td>2026</td></tr><tr><td><a href="https://docs.roboflow.com/models/supported-models/trocr.md">TrOCR</a></td><td><a href="https://docs.roboflow.com/models/model-licenses.md#other-models">Upstream license</a></td><td>2021</td></tr><tr><td><a href="https://docs.roboflow.com/models/supported-models/doctr.md">DocTR</a></td><td><a href="https://docs.roboflow.com/models/model-licenses.md#other-models">Upstream license</a></td><td>2025</td></tr><tr><td><a href="https://docs.roboflow.com/models/supported-models/easyocr.md">EasyOCR</a></td><td><a href="https://docs.roboflow.com/models/model-licenses.md#other-models">Upstream license</a></td><td>2024</td></tr></tbody></table>

## Improve OCR accuracy

The OCR models run pretrained, so you do not need to train one. To improve results, train an [Object Detection](https://docs.roboflow.com/models/supported-tasks/object-detection.md) model that finds the text region, and pass its crops to the OCR model.

## Deploy an OCR model

OCR models run pretrained. Call them through the [Serverless Cloud API](https://docs.roboflow.com/deployment/roboflow-cloud/serverless-api), or run them on your own hardware with [Roboflow Inference](https://docs.roboflow.com/deployment/self-hosted/self-hosted). See [Choosing a Deployment](https://docs.roboflow.com/deployment/choosing-a-deployment) to compare the options.
