T-Rex Label
T-Rex Label uses visual prompts and zero-shot detection to annotate similar objects across images without fine-tuning. Computer-vision teams can use it to prepare labeled datasets for complex scenes in a browser.
One of 375 tools in Image & Design
Who it's for
- Computer-vision engineers preparing datasets who want to use visual prompts to find repeated objects.
- Data scientists working on image projects involving rare objects or large batches of similar targets.
- Teams building datasets for the listed agriculture, retail, medical, logistics, or other visual-AI applications.
How it fits your workflow
Open the browser-based tool and provide a visual prompt, such as drawing a bounding box around an object. The page says the built-in detection model works without additional training or fine-tuning.
Use the prompt to detect similar objects in an image, then apply prompt information to other images for cross-image batch annotation.
Import and export mainstream data formats to connect the annotation work with a visual-AI workflow.
Pricing
Free: The page offers a “start for free” option but does not specify the scope or limits of a free tier.
The vendor doesn't publish prices. Check with T-Rex Label directly.
Prices checked on Oct 8, 2026 from the vendor's site. They can change; confirm before you buy.
Key features
- Zero-shot open-set detection
- The built-in detection model is described as identifying objects without additional model training or fine-tuning.
- Visual prompting
- Annotators can draw a bounding box around an object as a prompt, after which the model detects similar objects.
- Cross-image annotation
- A prompt can be applied to other images so similar targets can be annotated in batches.
- Smart segmentation and bounding boxes
- The page lists smart segmentation and smart bounding-box annotation among the tool's capabilities.
- AI pre-annotation
- AI-assisted pre-annotation is listed as a feature for preparing dataset labels.
- Browser-based access
- The tool runs in a browser and is described as requiring no installation or deployment.
Works with
- Data and model ecosystem references
- Machine-learning frameworks and services
Getting started
- Open T-Rex Label in a browser; the page says no installation or deployment is needed.
- Draw a bounding box around an object to provide a visual prompt.
- Use detection results and apply the prompt across other images for batch annotation.