"OpenCV" matching MCP tools:

• template-matchA

  VisionFlow MatchA

  Find where a small pattern image (a button, icon, logo or crop) appears inside a larger screenshot or picture, using OpenCV cv2.matchTemplate. Returns the best position, plus every non-overlapping position scoring above the threshold, as pixel boxes with scores. Use it for exact-looking pixel patterns at the same scale; for different scale, rotation or perspective use /v1/homography. Price: $0.003 a call (3 free calls a day without an API key).

  ConnectorNo auth

• feature-matchA

  VisionFlow MatchA

  Match local features between a pattern image and a larger image: detects ORB or SIFT keypoints in both, matches them with a brute-force matcher and Lowe's ratio test (OpenCV BFMatcher.knnMatch), and returns the good matches as point pairs sorted by distance. Use it to see which parts of a pattern are present and where, even when the pattern is scaled or rotated; for the pattern's outline use /v1/homography. Price: $0.004 a call (3 free calls a day without an API key).

  ConnectorNo auth

• shorti_scene_kitA

  ShortiA

  A served league video's generator scene kit: JSON, clay glTF, camera, light, beats, ledger buckets and C35 model adapters. A served league video's id or post link: reads its precompiled scene kit. The default summary keeps every source shot clock, keep/lift/free with evidence, export links, and each model's segment_count and first request (for_model=true). Read detail='full' for the unchanged complete legacy kit and all model requests. The full scene_kit.v0 has versioned OpenCV cameras, lens, trajectories, depth, light, beats, cuts and ledger buckets, a clay glTF export and four C35 adapters. Public identity is nr; the private twin joint uses only the owner's own likeness-consented assets, never league likeness. No analysis, generation, identity IO or spending runs in this request.

  ConnectorOAuth

• trace_from_image

  kernelcadA

  Use this when you need to trace features from a reference photo into waypoints. Trace pixel-space features from a reference photo into normalized [0..1] waypoints the agent can map to mm via a known scale anchor and feed to path().spline / path().nurbsSegment. Three backends are dispatched behind the scenes: `opencv` (deterministic; uniform-bg silhouette only), `vision-llm` (vision model; named points/cluttered backgrounds; needs ANTHROPIC_API_KEY or an OpenAI-compatible vision endpoint configured on the server), and `hybrid` (opencv silhouette + LLM-labeled named points). Default backend is `auto` — the tool picks based on the image's corner-color stddev. Accuracy honesty: opencv contour is geometrically exact; vision-LLM is typically 5–10% off on dense landmarks. Per-feature `confidence` is reported. Vision-LLM spend is billed to whoever configured the vision model. Pair with the `kernelcad-trace-from-image` skill for the conversion-to-mm pipeline. Hosted: give the file as an https:// URL, a data: URI (base64, max 25 MB) or "asset:<id>" from request_upload (when the file is on the user's device).

  ConnectorNo auth