Skip to content
cccskills
FreePublished on GitHub

panotovr

Create a Realsee VR space from local equirectangular panoramas using PanoToVR. Use for image-directory or manifest submission, resuming a known VR task, and retrieving its VR URL. For depth maps, point clouds, or camera poses use Argus instead.

The English summary is being prepared. Showing the original description.

See how to install

Files included (26)

  • SKILL.md3.7 KB
  • assets/manifest.example.json259 B
  • LICENSE17.0 KB
  • package-lock.json109.4 KB
  • package.json493 B
  • README.md1.1 KB
  • README.zh-CN.md1019 B
  • references/api-workflow.md1.0 KB
  • references/api-workflow.zh-CN.md997 B
  • references/host-differences.md986 B
  • references/host-differences.zh-CN.md855 B
  • references/input-modes.md1.4 KB
  • references/input-modes.zh-CN.md889 B
  • references/pano-to-vr.openapi.json14.1 KB
  • references/troubleshooting.md2.0 KB
  • references/troubleshooting.zh-CN.md1.2 KB
  • scripts/lib/api.mjs7.9 KB
  • scripts/lib/archive.mjs2.2 KB
  • scripts/lib/args.mjs3.9 KB
  • scripts/lib/config.mjs2.7 KB
  • scripts/lib/constants.mjs825 B
  • scripts/lib/input.mjs4.8 KB
  • scripts/lib/logger.mjs755 B
  • scripts/lib/state.mjs1.4 KB
  • scripts/run-panorama-to-vr.mjs6.1 KB
  • SKILL.zh-CN.md3.3 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

PanoToVR

English | 简体中文

PanoToVR uploads panoramas, creates a remote VR task, polls it, and writes result.json with status, task_code, project_id, vr_url, and workspace_dir. This is independent of Argus and local Blender modeling.

Before execution

  • Obtain consent for remote upload and VR task creation. Reuse explicit authorization for the same files and scope. Selecting files alone is not consent.
  • Resolve <skillDir> as the directory containing this file, regardless of the caller's working directory.
  • Require Node.js 22+ and npm 10+. Install locked dependencies with npm ci --prefix <skillDir> --omit=dev --ignore-scripts --no-audit --no-fund.
  • Configure REALSEE_APP_KEY, REALSEE_APP_SECRET, and REALSEE_REGION=global|cn locally. The runtime reads the process environment or .env in the current working directory; it also accepts CLAUDE_PLUGIN_OPTION_REALSEE_*. It does not load Argus's credential file. Never request secrets in chat or print them.
  • Confirm the account region; the runtime defaults to global when omitted. For missing access, use my.realsee.ai for global or my.realsee.cn for CN. Request the panorama-to-VR API capability through developer@realsee.com with account region, UserID, and IdentityID. Do not send an email without authorization.
  • Keep workspaces outside the installed skill and repository. State includes upload receipts and may contain temporary download URLs; treat it as private. The runtime clears proxy environment variables for its own process.

Execute

Image directory (auto-generated manifest):

node <skillDir>/scripts/run-panorama-to-vr.mjs --images-dir <imagesDir> --workspace <workspaceRoot> --json

Existing manifest (read input modes and use the synthetic template):

node <skillDir>/scripts/run-panorama-to-vr.mjs --manifest <manifestPath> --images-dir <imagesDir> --workspace <workspaceRoot> --json

Resume a known task without packaging, upload, or submission:

node <skillDir>/scripts/run-panorama-to-vr.mjs --workspace <workspaceRoot> --task-code <taskCode> --json

--workspace is the parent root, including on resume; the emitted workspace_dir is the task subdirectory. Optional --project-name, --poll-interval-ms, and --poll-max-attempts configure new runs/polling. Defaults are 30 seconds and 60 attempts; environment alternatives are REALSEE_POLL_INTERVAL_MS and REALSEE_POLL_MAX_ATTEMPTS.

Interpret and recover

Read result.json in the emitted workspace and report the task identity, status, workspace, and VR URL when present. A remote error may be returned as status: error even with exit code zero; inspect the JSON rather than relying only on exit code. Polling timeout does not prove remote failure. Resume a known task after interruption; if submission may have succeeded without returning a task code, stop and reconcile before submitting again. Do not claim a usable VR link when it is null.

The runtime normalizes JPEG/PNG/WebP to JPEG and checks manifest structure and file matches; it does not enforce Argus's 1–99/RGB8/exact-2:1 validation contract. Confirm equirectangular inputs rather than claiming those checks ran.

Load references only as needed: API workflow, host installation, troubleshooting.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

argus

Free

Use this skill to process one to 99 local exact 2:1 equirectangular panorama images with Realsee Argus, producing depth maps, a merged GLB point cloud, camera poses, optional intrinsics, and a validated local output index. Trigger for Argus panorama reconstruction, Argus ZIP input, or explicit Argus start/status/collect lifecycle requests. Do not trigger for panorama editing or stitching, arbitrary-photo 3D generation, existing GLB inspection, or research-only questions.

The English summary is being prepared. Showing the original description.

realsee-developer/skills10Updated Oct 20, 2026

Reconstruct an evidenced indoor space from local Realsee scan exports, point clouds, CAD, and panoramas as a separately editable native Blender scene. Use for scan-to-Blender space modeling, source-based geometry/material/light refinement, and requested walkthrough or physics exports. Argus panorama-to-depth inference is a separate capability.

The English summary is being prepared. Showing the original description.

realsee-developer/skills10Updated Oct 20, 2026

See all skills by realsee-developer

Report a problem with this skill