FAQ

Frequently Asked Questions

Below is a small list of the questions we commonly get about Rustici Generator. If you have your own question about Generator that isn’t answered here, please reach out to our team. We’re here to help if you have questions or need advice!

Email: support@rusticisoftware.com Phone: 866.497.2676




Questions:

What types of content are supported?

  • Rustici generator will work with content published from these authoring providers off the shelf:

    • Articulate Rise and Storyline
    • Adobe Captivate
    • dominKnow
    • Gomo Learning
    • Intellum Evolve
    • Easygenerator

    Generator will attempt to parse packaged content authored by other organizations.

  • Generator will also accept the following content types:

    • MP4 & MP3
    • VTT
    • PDF
    • URL

Does the AI part of Rustici Generator get trained to enhance the model further?

  • Generator’s language model is not trained with customers’ course content, but we do plan to support a “bring your own model” option for those who have special requirements or preferences.

Will Generator work with dispatch packages?

  • At this time, Rustici Generator is unable to parse through a dispatched eLearning course package. As you may know, the actual content is not stored within a dispatch package and only contains where to reference said content.

    However, our team is exploring options for how Rustici Generator could be used to help create more discoverability and usability of content within your library, which is a dispatch package.

Can Generator handle scanning content that is in other languages?

  • Support for non-English language depends on the model being used, but the default model (Anthropic Claude) supports a variety of languages. As for the other supported models, AI21 Labs Jurassic also has varying levels of support for multiple languages, while Amazon Titan Text is optimized for English.

How does Generator store and retain my content?

  • Unless otherwise specified, Generator data is retained indefinitely until explicitly deleted by a request from your system.
  • During an import job, content files are downloaded from the source location, parsed, then deleted.
  • Optionally, imported artifacts are also saved to an S3 bucket for easier file identification and troubleshooting.
    • This feature is enabled by default for managed hosting instances of Generator, and the content expires after 7 days.
    • For self-hosted instances of Generator (see HOSTING__CONTENT_IMPORT_S3_CACHE_TAG in the hosting variables page), we recommend configuring the bucket with a lifecycle rule that automatically deletes content from the bucket after a certain amount of time. Without a lifecycle rule, the imported content will never be deleted.
  • Once imported, the extracted text is saved in a few places:
    • The raw text is stored in the database
    • Content types that rely on transcription (videos and audio) will have a VTT copy of the subtitles saved to S3.
    • To support lexical search, the text is saved as vectors in the database
    • To support semantic search, the text is embedded and saved as vectors in the database
  • When a content version is deleted, all information about that version is also deleted. This includes extracted text, vectors derived from that text, fields generated for that content version, etc.
  • When a tenant’s data is deleted using a tenant deletion job, then all data contained within that tenant (config values, previous jobs, imported content, etc) is deleted.