Overview
T-Rex Label is an online, browser-first image and video annotation platform designed to accelerate computer vision dataset creation. It emphasizes zero-setup workflows and model-assisted annotation: users provide visual prompts (for example, draw a single bounding box) and built-in open-set detection models (Grounding DINO, DINO-X and the platform’s T‑Rex2) automatically detect and propagate labels across images and batches without additional fine-tuning. The product supports common output formats (COCO, YOLO), offers mask and bounding-box annotations, and provides AI pre-annotation, cross-image inference, and batch annotation features to scale labeling tasks quickly. T-Rex Label targets fast onboarding, minimal maintenance, and applicability across industries such as agriculture, logistics, medicine, and manufacturing.
Key Features
Zero-shot/Open-set Detection
Built-in detectors (Grounding DINO, DINO-X, T‑Rex2) accept visual/text prompts to detect novel objects without retraining.
Cross-image Annotation (Batch Propagation)
Draw a prompt in one image and propagate labels automatically across similar scenes and entire image batches.
AI Pre-annotation (Mask & Bounding Box)
Automatic mask and bounding-box predictions to accelerate manual correction and reduce labeling time.
Browser-first Workspace
No installation required; users import images/videos via drag-and-drop and annotate directly in the browser workspace.
Dataset Export & Format Support
Exports compatible with COCO and YOLO formats to integrate with common training pipelines.
Multimodal Visual Prompting (T‑Rex2)
Accepts visual prompts and excels at recognizing items outside initial training sets, enabling annotation of rare objects and complex scenes.
Who Can Use This Tool?
- Researchers:Rapidly create labeled datasets and iterate experiments with AI-assisted annotation.
- ML Engineers:Prepare COCO/YOLO datasets and scale annotations using zero-shot detection and batch propagation.
- Data Labelers:Speed up manual labeling with AI pre-annotations and cross-image propagation for large projects.
- Enterprises:Deploy fast annotation workflows for industry use cases like agriculture, logistics, and medical imaging.
Pricing Plans
Free tier to try browser-based AI annotation tool with core features.
- ✓Browser-based, zero-setup annotation
- ✓AI Mask
- ✓AI Bounding Box
- ✓Cross Image annotation
- ✓AI Pre-annotation
- ✓One-click detection without fine-tuning
- ✓Batch annotation via visual prompts
- ✓Support for popular dataset formats (COCO, YOLO)
Pros & Cons
✓ Pros
- ✓Zero-setup, browser-based (no install) and fast onboarding
- ✓Built-in open-set / zero-shot detection (Grounding DINO, DINO-X, T‑Rex2) for one-click pre-annotation
- ✓Cross-image inference and batch annotation via visual prompts for scalable workflows
- ✓Exports to COCO/YOLO formats, covering common CV training pipelines
- ✓Supports both image and video import (drag-and-drop media preparation)
✗ Cons
- ✗Limited public third-party reviews and community feedback
- ✗No clearly visible paid/enterprise pricing page or plan tiers discovered
- ✗Mobile workspace not supported (desktop/browser recommended)
- ✗Detailed docs, API endpoints, or enterprise SLAs not clearly found on public pages