Optimization Settings

Optimization settings define how an avatar is processed and prepared for use in your virtual experience. Optimization Settings can be set up from the MSquared Dashboard.

Processing presets

Currently, you may select the following processing presets:

• Low

• Medium

• High

The presets control the texture sizes used and the level of detail of the processed mode. More fine-grained control for processing will be available in the future.

Processing duration

Processing an avatar is expected to take between 5 and 15 seconds. During busy periods it may take longer. Furthermore, if you receive a 429 it is recommended to retry using exponential backoff.

Processing metering

Processing is a metered operation. Therefore, we recommend only allowing your backend to call the processing endpoint, not handling this directly from client machines, which would expose your MSquared API key. Within a project, if you request to process the same avatar multiple times with the same optimization settings in a 12-month period you will not be charged multiple times. We cache previously processed models to save you time and money.

What to do with processing results

The response from the processing endpoint will include two publicly downloadable URLs, one for the raw glb and an MML file, which wraps the glb ready to use in MSquared experiences. As the call will have been made from your backend it is up to you to decide what to do with these. For example you could return the download URLs to the client's machine if you want them to have direct access to the file or you could attach the model URL to the calling user's profile. Model's will be stored in our buckets for at least 12 months - however after that they may be cleared - therefore it is recommended you transfer the model to your own storage if you need storage lasting longer than this.