Live demo
Locate text in your image.
Drop a production image, label, sign, or document. The image is processed in memory and returned with detected text regions. This demo locates text; recognition will be added separately.
Images and regions
JPG, PNG, WebP, optional ROI and orientation handling in the planned API.
Text + geometry
Recognized strings, confidence, bounding geometry, ordering, and visual overlays.
API or offline SDK
Online evaluation first, then a Windows SDK or hardware-specific optimization.
Publication standard
The live timing is specific to this request and server. Accuracy claims will only be published after the evaluation set, metric definition, hardware, preprocessing, and software version are documented.
Current integration
- FAST text localization through OpenVINO
- Per-region confidence and polygon geometry
- Visual overlay returned by the inference server
- Separate preprocessing, inference, and postprocessing timing