Upload, annotate, manage, and version image and video datasets on Roboflow.
Everything for building high-quality datasets: uploading data, labeling it, organizing projects, and creating versioned snapshots for training.
Create a Project
Object detection, classification, segmentation, keypoints, and more.
Upload Data
Drag-and-drop, API, or cloud import from S3, GCS, and Azure.
Upload a Dataset
Bring in an existing labeled dataset in any common format.
Download a Dataset
Export images and annotations in dozens of formats.
Roboflow Annotate
Manual tools, keyboard shortcuts, and team collaboration.
AI Labeling
Label Assist, Smart Select, Box Prompting, and Auto Label.
Manage Datasets
Organize batches, search, classes, and tags.
Manage Images
Image-level operations via the API and SDK.
Project Folders
Group related projects together into folders.
Dataset Versions
Preprocessing, augmentation, and versioned exports.
Dataset Health Check
Class balance and image dimension insights.
Project operations: Merge · Rename · Delete
Roboflow Universe
Find, explore, and fork 200k+ public datasets and models.
Once a dataset version is ready, head to Models to train a model on it.
Last updated 16 days ago
Was this helpful?
