T-Rex Label is a browser annotation tool built around visual prompts. You draw a box around an object, and the site says the tool finds similar objects and applies that prompt to other images. It runs in the browser, so there's nothing to install or deploy.
Checked against the site's own pages on .
Our take
Worth a look if your labeling queue is full of repeated objects spread across many images. The cross-image prompt is the part that saves time, since one box can carry over to the rest of the batch. We'd still check the export formats against your own pipeline before committing a team to it.
Best for
- Computer vision engineers labeling repeated objects across image sets
- Teams that want a browser tool without installation or deployment
- Annotation work on objects outside a model's training classes
Sources
- trexlabel.com/ (opens in a new tab) Product description, prompt workflow and browser-based delivery