Run local SAM3 image segmentation and video prediction APIs.
npx skills add https://github.com/EurecaMoment/BenchClaw --skill sam3-local
Please help me install this Agent Skill. Skill: sam3-local Source: https://github.com/EurecaMoment/BenchClaw/tree/main/BenchClaw/annotation-tool/sam3 Command: npx skills add https://github.com/EurecaMoment/BenchClaw --skill sam3-local
This Skill provides a local interface to the SAM3 annotation tool, enabling text-grounded image segmentation, box-prompted image segmentation, candidate mask generation, and direct access to the SAM3 video predictor API.
Run the local SAM3 annotation tool on the image 'example.png' with the text prompt 'car' and save masks to 'output_masks'.
High-intent search queries and answers about installing and using this skill.
Local image segmentation with the SAM3 annotation tool is performed using text-grounded or box-prompted inputs to generate candidate masks. This process operates as a local service, allowing you to process images like 'example.png' directly without reloading the model for each new prompt.
Yes, you can access the SAM3 video predictor API directly to run local video tracking sessions. The service is designed to maintain the loaded model in memory, eliminating the need for additional model reloading between consecutive video prediction tasks.
Text-grounded image segmentation is the process of isolating objects within an image based on natural language descriptions. You need it for object detection tasks when you want to automatically identify and segment specific entities, such as a 'car', using text prompts instead of manual pixel selection.
Yes, a running local SAM3 service is required to use box-prompted segmentation. This Skill functions as a local interface to the SAM3 annotation tool and relies entirely on the local environment to process bounding box inputs and generate corresponding candidate masks.
The best way to generate candidate masks for local video prediction is by running a SAM3 video tracking session through the local service. This approach provides direct access to the SAM3 video predictor API, enabling efficient mask generation and tracking across video frames.
tordksProvide a Python API for promptable SAM3 segmentation on images and videos.
jayll1303Train and export Ultralytics YOLO models for detection and tracking.
automatedigitalSegment arbitrary objects in images using point, box, or mask prompts.
EurecaMomentOrchestrate VLM, YOLOE, SAM3, and Depth Anything 3 to generate image annotations.
