Realsee Argus Skill 2.0.0

Realsee Argus Skill 2.0

Realsee Argus is a world-leading 3D vision foundation model. From a photo, panorama, or sparse views, it reconstructs metric 3D structure in milliseconds: pose, depth, point clouds, and renderable geometry.

Public evidence includes a 1.31B-parameter model and the Realsee3D benchmark's 10K complete indoor scenes, 95,962 room units, and 299,073 panoramic viewpoints.

Realsee Skills brings Argus to agent and CLI workflows. Skill 2.0 has a deliberately narrower public contract: it processes 1–99 local RGB8 panoramas with exact 2:1 dimensions and produces EXR depth maps, one merged GLB point cloud, per-image camera poses, optional intrinsics, and a validated local result index.

Install

npx skills add realsee-developer/skills --skill argus --agent codex
npx skills add realsee-developer/skills --skill argus --agent claude-code

Claude Code marketplace users can also install realsee-skills@realsee-developer-skills.

Release highlights

• Supports both Global and CN Realsee Gateway regions through the same Skill contract.
• Adds explicit start, status, and collect lifecycle commands.
• Provides verified first-party example manifests and a CDN downloader without distributing panorama JPEGs in Git.
• Validates depth maps, merged GLB output, camera poses, optional intrinsics, and the local result index.
• Removes the legacy single-image VGGT fallback; pin v1.0.2 for the 1.x square-input workflow.

Argus runs upload selected inputs to Realsee remote services and require explicit user consent.

Usage guide · Install guides · 2.0 migration · Argus product

Changes

• Argus 2.0 stable release: #4
• Full changelog