T-Rex Label

T-Rex Label is a browser-based data annotation platform for B2B teams that need to build computer vision training datasets across industries such as healthcare, automotive, robotics, agriculture, and logistics.

Key capabilities include:

• AI-assisted labeling and pre-annotation using built-in visual models (including Grounding DINO, DINO-X, and T-Rex2) with no fine-tuning required
• 2D bounding box and mask (segmentation) annotations for common detection and segmentation workflows
• Visual prompt-based object detection, where a user draws a bounding box to find similar objects in an image, plus cross-image inference to apply prompts across multiple images
• Dataset format compatibility with COCO and YOLO import and export for downstream training pipelines

Pricing: Undisclosed, contact for pricing.

What makes it notable is its focus on one-step, in-browser prompting paired with open-set, zero-shot style detection in T-Rex2, designed to recognize objects beyond the initial training set without retraining, which can reduce manual labeling effort on rare or long-tail classes.