Skip to main content
Free Browser ToolNo installation needed

YOLO TXT to COCO JSON Converter

Merge individual YOLO .txt label files or a ZIP archive into a single, standardized COCO JSON dataset file. Restores absolute pixel coordinates and builds a clean category mapping for PyTorch, Detectron2, and MMDetection.

Convert annotation format
Upload one annotation file and download it in a different format.

One file only

COCO JSON, YOLO TXT (ZIP), Pascal VOC XML, CVAT XML, LabelMe JSON, CSV, TSV, or JSONL.

Quick conversions

Converting YOLO Normalized Labels to COCO JSON

YOLO Format (Per Image .txt)

YOLO stores one txt file per image with normalized values:

0 0.500000 0.500000 0.250000 0.400000
1 0.200000 0.800000 0.100000 0.150000

COCO JSON Format (Unified)

COCO combines all images, categories, and bboxes into one JSON:

{
  "images": [{"id": 1, "file_name": "img1.jpg", "width": 1920, "height": 1080}],
  "categories": [{"id": 0, "name": "person"}],
  "annotations": [{"id": 1, "image_id": 1, "category_id": 0, "bbox": [720, 324, 480, 432]}]
}

Key Benefits of LabelOp In-Browser Converter

  • Zero Server Uploads: Safe for proprietary, medical, and private datasets.
  • Detectron2 & MMDetection Ready: Clean COCO standard structures.
  • ZIP Batch Support: Drag and drop your entire labels directory at once.
  • Instant Download: Converted in seconds directly in WebAssembly/JS.

Want an end-to-end computer vision workspace?

LabelOp provides automated labeling with SAM 2, team review queues, and model training.

Get Started Free