Version 1.2.0 — Now Available

Annotation at the
Speed of AI

Professional desktop labeling tool for object detection, segmentation, pose estimation, face detection and classification — with built-in AI auto-labeling.

Buy SwiftLabel Download Trial
26
Export Formats
6
Project Types
5
AI Pipelines
Native
C++ Performance
SwiftLabel instance segmentation project with a polygon mask, class panel and dataset statistics
Features

Everything you need to
build great datasets

A complete annotation workflow in a single native desktop application — no cloud, no subscription.

🏷️

5 Annotation Modes

Bounding boxes, polygons, keypoints, face landmarks and image-level classification — all in one tool with dedicated UI per project type.

⚡

AI Auto-Labeling

Run object detection, instance segmentation, body & hand pose or face landmark models on your dataset.

📦

26 Export Formats

YOLO v3–v12, COCO, Pascal VOC, LabelMe, CreateML, CVAT, Darknet, Florence-2 and more — export ready to train.

🎬

Video Annotation

Import MP4 / AVI and annotate frame by frame with a built-in scrubber. Annotations persist per frame across sessions.

📊

Dataset Statistics

Live overlay showing labeled progress, annotation count, and per-class distribution charts — always visible while you work.

🔒

Fully Offline

100% local processing. Your images and annotations never leave your machine. No internet connection required.

Product Tour

Real annotation workflows, not mockups

SwiftLabel keeps the canvas, classes, annotations, files, frame navigation and dataset statistics visible while you work.

Instance Segmentation

Draw and review precise masks

Polygon tools, class colors and live annotation lists stay in the same workspace, so mask cleanup remains fast even on detailed objects.

Polygon masks Class colors Dataset stats
SwiftLabel polygon mask editing workflow
Video Object Detection

Annotate frame by frame

Import video, move through frames with the scrubber, and keep bounding boxes tied to the selected frame and class list.

Video frames Bounding boxes Playback controls
SwiftLabel video object detection workflow with player bounding boxes
Classification Dataset

Keep dataset progress visible

File lists, per-class totals and labeling coverage stay in view, helping teams spot gaps before export.

280 images 100% labeled Class totals
SwiftLabel classification dataset progress with per-class statistics
Project Types

Built for every computer vision task

Each project type activates the right tools and export formats automatically — no manual configuration needed.

🔲

Object Detection

Draw bounding boxes around objects with class labels. Supports YOLO, COCO, VOC and more export formats.

BBox AI Auto-Label
🖌️

Semantic Segmentation

Pixel-level polygon masks where every region gets a class label. Ideal for scene understanding tasks.

Polygon SAM Click-to-Segment
🔷

Instance Segmentation

Individual polygon masks per object instance with class labels and per-instance IDs.

Polygon AI Auto-Label SAM
🦾

Pose Estimation

COCO 17 body and 21-point hand skeletons with AI assistance.

Keypoints AI Body AI Hand
🙂

Face Detection

Face bounding boxes with 5-point landmark detection.

BBox AI Face
🏷️

Classification

Assign a single class label to each image in the dataset.

Select Multiclass
AI Tools

Let AI do the heavy lifting

Load ONNX models from your local models folder and pre-label datasets fully offline.

⚡ Detect

Object Detection

Draw bounding boxes around objects with class labels. Run YOLOv8/YOLO11 ONNX models for AI auto-labeling across your entire dataset.

🖌️ Semantic

Semantic Segmentation

Pixel-level polygon masks where every region gets a class label. Use SAM click-to-segment for instant zero-shot masks on any object.

✂️ Instance

Instance Segmentation

Individual polygon masks per object instance. YOLO segmentation models and SAM generate masks automatically — review and approve with one keystroke.

🦾 Pose

Pose Estimation

COCO 17-keypoint body skeletons and 21-keypoint hand skeletons with per-finger coloring. Switch between Body and Hand models in one dialog.

🙂 Face

Face Detection

Face bounding boxes with 5-point landmarks (eyes, nose, mouth corners). Batch-run across all images and approve results.

🏷️ Classify

Classification

Assign a single class label to each image in the dataset. Supports multiclass labeling with quick keyboard-driven selection.

Export Formats

Train with any framework

26 export formats covering major training workflows — from vanilla YOLO to Apple CreateML.

YOLOv8 COCO JSON Pascal VOC LabelMe CreateML YOLOv5 PyTorch YOLOv9 Darknet YOLO11 CVAT XML Florence-2 Roboflow YOLO YOLOv7 RetinaNet Multiclass PaliGemma SwiftPackage YOLOv12 YOLO Pose YOLOv8 COCO JSON Pascal VOC LabelMe CreateML YOLOv5 PyTorch YOLOv9 Darknet YOLO11 CVAT XML Florence-2 Roboflow YOLO YOLOv7 RetinaNet Multiclass PaliGemma SwiftPackage YOLOv12 YOLO Pose
Shortcuts

Designed for speed

Every action has a keyboard shortcut so you never have to reach for the mouse.

Tools

Select / Move
V
Bounding Box
B
Polygon
P
Keypoint
K
Classification
C

Navigation

Next image
D
Previous image
A
Import dataset
F5
Video play / pause
Space
Fit to window
F

Editing

Undo
CtrlZ
Redo
CtrlY
Save project
CtrlS
Toggle labels
L
Delete selected
Del

View

Zoom in
+
Zoom out
−
100% zoom
G
Show shortcuts
F2
Pricing

One-time payment.
No subscription.

Try everything free for 14 days — no account, no credit card. Buy once, keep it forever.

Trial

Free

Full app, every feature, for 14 days. No account or credit card required.

All features 14 days

Lifetime License

From $29

One-time payment — use forever. Includes use on up to 2 devices for one user, so your workstation and laptop are both covered.

Lifetime updates 1 user / 2 devices
Buy SwiftLabel

Licenses are delivered instantly by email after checkout. USD international pricing and TRY pricing for Turkish cards are available at checkout.

Buying for a team? See enterprise procurement, security, and support details.

One user, two devices Use your lifetime license on your main workstation and one additional machine.
Private datasets Annotation work runs locally; your images are not uploaded to SwiftLabel.
Secure checkout Card payment is handled by iyzico and the license key is delivered by email.
Recoverable order After purchase, the private order link shows payment, license and email status.
Enterprise procurement Security and privacy Support policy System status Release notes
Download

Start annotating
in minutes

One installer, no cloud account, no subscription. Your data stays on your machine.

Download v1.2.0
14-day free trial Windows & Linux · 64-bit No internet required ONNX model support