FILE HANDLING

Your model.
Your decision.

When you choose a file, the preview runs in your browser. Clicking Optimize uploads the model to our server for processing with gltfpack.

Temporary processing

Input and scratch files are removed when processing completes or fails. Successful output is available until 60 minutes after upload. The cleanup worker removes expired files, normally within a few seconds; cleanup resumes after service outages. Interrupted uploads are cleaned after five minutes. Task metadata expires with the task.

Access and limits

Each task has a random access token. Keep it private. We accept one embedded-resource GLB up to 50 MB, process for up to 120 seconds, and limit output to 100 MB. Files are not published in a gallery or used for training.

Operational data

We keep aggregate daily visit, accepted upload, task success/failure and result-fetch counts for up to 90 days. They do not contain model contents or original filenames. Temporary hashed network identifiers are used for rate limiting and expire after their rate window. Hosting infrastructure may maintain separate access logs; deployers must configure their retention. No third-party analytics or remote viewer assets are loaded by this application.

Check before using the result

Optimization may change scene structure, precision and animations. Custom extensions may be discarded. Keep your original file and verify the output in your target application. This is an independently operated interface to the MIT-licensed gltfpack engine.