Commercial Dataset Annotation Suite

Label your footage.
Train your own model.

Built for high-frame-rate capture cards and live video — not generic static photo sets. Segment Anything 2 (SAM 2.1), temporal propagation across video clips, open-vocabulary text prompting, and on-device training running 100% on your machine.

Start for $15/mo → Explore all features
Runs 100% locally • Your images and datasets never leave your machine • Cancel any time
Interactive AI Studio SentinelLabeler Canvas Preview
Click anywhere on the canvas to place a SAM 2 segmentation prompt.
Dataset Telemetry
TOTAL FRAMES
3,420
CLASSES
4 Classes
FORMAT
YOLO11 / ONNX
DATA PRIVACY
100% Local
Pro Tip: Press Space to auto-annotate or P to propagate across 300 sequential frames.
The Workflow

From raw capture card to trained model

A seamless 4-step pipeline designed specifically for high-speed game vision.

1

Capture Ingest

Stream directly from Elgato HD60 X, Magewell, AverMedia, or local video files. Smart deduplication automatically filters identical idle frames.

2

AI-Assisted Labeling

Segment Anything 2 (SAM 2.1) click-to-mask, open-vocabulary zero-shot text prompting with YOLO-World, and SAHI tiled inference for distant micro-targets.

3

Temporal Propagation

Annotate a single frame, then propagate boxes and masks forward across up to 300 video frames automatically using optical flow.

4

On-Device Training

Train YOLOv8, YOLOv9, or YOLO11 directly on your local NVIDIA GPU with automatic isolated Python environment — or export straight to Sentinel Core Vision.

Why SentinelLabeler

Engineered for Video. Unmatched for Local Privacy.

Compare SentinelLabeler with generic photo-labeling web tools.

Capability SentinelLabeler Roboflow CVAT (Cloud) LabelImg
100% On-Device Privacy
Footage & weights never upload		 ✓ 100% Local Cloud Upload Cloud / Server ✓ Local
Live Capture Card Stream Ingest
Elgato, Magewell, AverMedia		 ✓ Native DirectShow ✕ Photos Only ✕ Video Upload ✕ Photos Only
Temporal Video Propagation
Label 1 frame, track whole clip		 ✓ Built-in (300 frames) ✕ Manual Partial (Server Tracker) ✕ None
Segment Anything 2 (SAM 2.1)
Instant polygon mask generation		 ✓ Local GPU SAM 2 Cloud Paid Tier External Server ✕ None
Dual-Model Live A/B Compare
Side-by-side live video testing		 ✓ Side-by-Side ✕ None ✕ None ✕ None
Isolated Self-Installing Python
Zero CUDA or PATH headaches		 ✓ Fully Automated N/A (Web App) Complex Docker Manual Python
Direct Sentinel Core Vision Link
1,000Hz motor loop export		 ✓ 1-Click Hot Swap ✕ None ✕ None ✕ None
Monthly Pricing $15 / mo (Unlimited) $249+ / mo (Cloud limits) Free to $33/mo Free (Discontinued)
Simple Pricing

Unlimited labeling. Zero cloud limits.

Cancel any time. Every dataset and model you export remains yours forever, even if your subscription ends.

Monthly Plan
Full access, billed month to month. Cancel whenever.
$15/ month
  • Unlimited dataset sizes & frames
  • SAM 2.1 & YOLO-World auto-annotation
  • Temporal video box propagation
  • Dual-model live compare
  • Local NVIDIA CUDA GPU training
  • YOLOv8, YOLOv9, YOLO11 & ONNX export
Start Monthly →
2 Months Free
Annual Plan
Full access, billed once per year. Best value for creators.
$120/ year
  • Everything included in Monthly
  • Equivalent to just $10 / month
  • Priority updates & model architecture support
  • Priority support on Discord
Get Annual Plan →
Secure Stripe checkout Cancel anytime with 1 click Runs 100% on your machine