sam2_process_video – 1 MCP server offers this tool
What it usually takes
async_frame_load– in 1 of 1drop_masks– in 1 of 1every_x– in 1 of 1masks– in 1 of 1ref_frame_idx– in 1 of 1sample_fps– in 1 of 1variant– in 1 of 1video_load_device– in 1 of 1video_path– in 1 of 1
No server marks it as only reading: treat it as able to change data.
From each publisher's own tool list. We never call a tool.
Servers
| Server | Inputs | Read-only? |
|---|---|---|
Sam2 co.huggingface.gf-john/sam2 SAM2 video segmentation: track objects through a video given their segmentation masks on a reference frame. Seed the objects to track by supplying their masks on the reference frame (ref_frame_idx); SAM2 then propagates each object forward -- and, when ref_frame_idx is nonzero, also backward -- through every frame. The video can optionally be downsampled before tracking with sample_fps or every_x; all frame indices (the "frame" field of every detection AND the ref_frame_idx input) are 0-based positions in the SAMPLED frame sequence, NOT original video frame numbers, so with every_x=5 the detection "frame": 2 corresponds to original frame 10. Returns a JSON list of per-frame detections, one entry per tracked object per frame in which it appears (frames where an object is absent are skipped). Each detection is a dict with: "frame" (integer sampled-frame index as defined above); "track_id" (integer object id, stable across frames); "x", "y", "w", "h" (the object's bounding box as top-left | async_frame_load,drop_masks,every_x,masks,ref_frame_idx,sample_fps,variant,video_load_device,video_path | Not stated |