Skip to main content

API & Integrations

The SuperSplat API lets applications publish Gaussian splats to superspl.at, list the scenes owned by the authenticated user, look up any scene by id, and download scenes whose owners allow it. Use it to connect a capture application, training pipeline, conversion tool, or internal content workflow directly to SuperSplat.

Uploaded scenes appear on the user's Manage page, where they can be edited, opened in Studio, or shared through their public scene page.

Complete API reference

For endpoint definitions, request schemas, and response examples, see the SuperSplat API reference.

When to use the API​

The API is useful when you want to:

  • Publish splats without opening the SuperSplat Editor.
  • Add SuperSplat publishing to a capture or reconstruction application.
  • Batch-upload output from a training or conversion pipeline.
  • List the scenes owned by a user.
  • Look up any scene by id, and import downloadable scenes into your own tools.
  • Resume a large upload after an interrupted connection.

Authentication​

All API operations require a PlayCanvas access token. See REST API authorization to create a token, then send it in the Authorization header:

Authorization: Bearer YOUR_ACCESS_TOKEN

Treat access tokens as secrets. Do not embed them in public client-side code or share them with another user.

Resumable upload flow​

SuperSplat uses multipart uploads so clients can reliably transfer large scene files:

  1. Create an upload session with the source format and optional scene metadata.
  2. Request signed upload URLs for one or more file parts.
  3. Upload those parts directly to storage.
  4. Complete the upload using the returned part ETags.

SuperSplat creates the scene after the upload is completed. The scene may remain in processing while conversion and optimization finish. New scenes are unlisted by default.

Downloading scenes​

GET /v1/splats/{splatId} returns any scene by id, whoever owns it, listed or unlisted. When its downloadable field is true, its owner allows downloads under a Creative Commons license (given in license), and content describes its published files. A download is the only way the API gives you a scene's files:

  1. Request a download with POST /v1/splats/{splatId}/downloads, giving a target format the scene offers.
  2. Poll GET /v1/splats/{splatId}/downloads/{target} until its status is ready. If it is failed, request the download again.
  3. Get a short-lived URL with POST /v1/splats/{splatId}/downloads/{target}/file, and start fetching it before urlExpiresAt.
targetOffered forContents
plyEvery downloadable sceneA conversion to a standard PLY
sogScenes published as SOGThe published SOG files, unchanged
ssogScenes published as Streamed SOGThe published Streamed SOG files with their levels of detail, unchanged

Each download is a ZIP archive with the files at its root and a license.txt giving the scene's title, author, source and license. Downloads are shared and cached, so a scene someone downloaded recently is often ready straight away. A download lasts while the scene stays the same. Once the scene is republished, renamed or relicensed, the download returns 404 and you request it again.

Requesting and polling a download are free. Getting the file counts towards the account's daily download allowance, once per scene per UTC day.

Limits​

When a request goes over a limit, the API answers 429 with a code that says which. A 429 without a code means the client is sending requests too quickly.

codeMeaning
daily_publish_limit_reachedThe account has used today's publishing allowance, measured in bytes of compressed output. It resets at 00:00 UTC.
too_many_processingThe account already has as many scenes uploading or processing as it may at once.
publish_in_progressAnother publish for the account is starting.
daily_download_limit_reachedThe account has downloaded as many scenes as it may today (UTC).

An upload larger than its source format allows is rejected with the largest size allowed. The API reference gives the fields of each error.

Client integrations​

The same API supports both end-user integrations and custom automation:

Clients can identify themselves through the optional uploadClient field:

{
"uploadClient": {
"id": "example-client",
"version": "1.0.0"
}
}

Use uploadClient for the application performing the upload. Use softwareTools separately to record the tools used to create the scene.

See also​