PaddleOCR Text Recognition

AI Platform & Data PaddleOCR 3.7.0 Hardened AMI

PaddleOCR 3.7.0 on Ubuntu 24.04 reads English text from PNG and JPEG images on CPU. Package 0.1.3 returns each word with a confidence score and a bounding box. A browser page draws the boxes on your image.

Get started View on AWS Marketplace

AMI release

0.1.3 (PaddleOCR 3.7.0)

Recommended

t3.large

Storage

32 GiB+ encrypted gp3

Access

Web UI in your browser

Released

Oct 24, 2026

Delivery

Amazon Machine Image in your own AWS account

Extract English text from PNG and JPEG images on CPU.

JSON output with each word, a confidence score and a bounding box; a browser page shows the boxes on your image.

Private HTTPS on this instance with a per-instance password. Images stay in your AWS account.

Quick start

Full guide →

1. 1 Subscribe
2. 2 Launch the instance
3. 3 Connect and run the walkthrough
4. 4 Sign in
5. 5 Read an image in the shell

What is on the image

This guide matches Marketplace package 0.1.3 (PaddleOCR 3.7.0). Check the listing for the version currently on offer.

• PaddleOCR 3.7.0 on CPU, for PNG and JPEG.
• HTTPS on port 8443, with a per-instance username and password in ~/.config/sushi/paddleocr/access.json.
• A command, sushi-paddleocr-api ocr page.png, which prints the text.
• A walkthrough, /opt/sushi/docs/demo.sh, and notes at /opt/sushi/docs/README.md.
• Stored jobs to back up before you replace the instance.

Images stay on the instance. The listing does not describe sending them to an external OCR service.

Recommended instance

t3.large with at least 32 GiB of encrypted gp3 root storage and IMDSv2. Leave inbound application ports closed. Session Manager does not need a key pair or an inbound rule.

Billing

The listing states a software rate of USD 0.03 per instance-hour. Confirm the offer before you subscribe. EC2, EBS, and data transfer are billed separately.

Tutorial

Watch on YouTube.

Where to go next

• Getting started: launch, sign in, and read an image.
• Configuration: the account file and the command.
• Operations: stored jobs and support.
• Troubleshooting: the tunnel and the file type.
• Software Sushi: the company site.

Upstream project

PaddleOCR is the open source recognizer. Software Sushi packages this image and does not claim upstream or AWS endorsement. Use /opt/sushi/docs/README.md for this build. The listing does not give a separate upstream URL.

Last updated November 1, 2026 · Suggest an improvement