Skip to content
Open Source · MIT Licensed

YOLO's NDJSON
Converter

Convert Ultralytics NDJSON and JSONL exports to YOLO, COCO JSON, and Pascal VOC formats. Fast, private, and runs entirely on your machine.

Terminal
$curl -fsSL https://yolondjson.zip/install.sh | sh
or

Available on macOS, Windows, and Linux · ~5 MB

Having trouble? Check the troubleshooting guide

Supports:DetectionSegmentationPoseOBBClassification

Why I Built This

A small companion tool for Ultralytics users

💙

Built on Love for Ultralytics

Ultralytics Platform is genuinely one of the best tools for vision AI - fast, intuitive, and constantly pushing boundaries. I use it daily and love what they're building for the community.

🔧

Bridging the Gap

Their NDJSON export works great inside the Ultralytics ecosystem. But when I wanted to train locally, try other frameworks, or simply browse my dataset as organized folders - the manual conversion added up. So I built this tool to do it all in one click.

This isn't a replacement - it's a small helper that extends your workflow beyond the cloud. If it saves you even a little time, I'll consider it a win.

Supported Formats

Convert NDJSON to COCO JSON, Pascal VOC, and all YOLO versions for local training.

YOLO26 (latest)

TXT annotations and YAML config used with YOLO26.

DetectionSegmentationPoseOBBClassification

YOLOv12

TXT annotations and YAML config used with YOLOv12.

DetectionSegmentationPoseOBBClassification

YOLO11

TXT annotations and YAML config used with YOLO11.

DetectionSegmentationPoseOBBClassification

YOLOv9

TXT annotations and YAML config used with YOLOv9.

DetectionSegmentation

YOLOv8

TXT annotations and YAML config used with YOLOv8.

DetectionSegmentationPoseOBBClassification

YOLOv7

TXT annotations and YAML config used with YOLOv7.

Detection

YOLOv5

TXT annotations and YAML config used with YOLOv5.

DetectionSegmentationClassification

YOLO Darknet

Darknet TXT annotations used with YOLO Darknet (both v3 and v4) and YOLOv3 PyTorch.

DetectionClassification

COCO JSON

COCO JSON annotations are used with EfficientDet Pytorch and Detectron 2.

DetectionSegmentationPose

Pascal VOC

Common XML annotation format for local data munging (pioneered by ImageNet).

DetectionSegmentationClassification

TFRecord

Coming Soon

TFRecord binary format used for both Tensorflow 1.5 and Tensorflow 2.0 Object Detection models.

Detection

CreateML

Coming Soon

CreateML JSON format is used with Apple's CreateML and Turi Create tools.

DetectionClassification

How to Convert Ultralytics NDJSON to YOLO

Three steps. One ZIP file. Done.

01

Select Your NDJSON

Pick your .ndjson or .jsonl export file from your YOLO's project.

02

Choose a Format

Select from 12 target formats including YOLO's, COCO, Pascal VOC, and more.

03

Download Your ZIP

Get a complete ZIP with converted annotations, images, and config files.

Built for Speed & Privacy

For ML engineers who need to move fast between frameworks.

Fast Downloads

100 concurrent connections download your images in parallel. Built with Rust and Tokio for maximum throughput.

12 Formats, 5 Task Types

From YOLO's to COCO JSON to Pascal VOC. Supports detection, segmentation, pose estimation, OBB, and classification.

100% Private

Everything runs on your machine. No data is ever sent to external servers. Your annotations stay yours.

Frequently Asked Questions

Common questions about converting Ultralytics NDJSON exports

Like This Tool?

If this tool helped you, consider giving it a star on GitHub or contributing to make it even better.