Skip to main content

​
What the Content Library does

The Content Library is the source-material layer for LearnTerms. It is where admins and contributors upload documents, organize them at the cohort level, and turn them into chunks that can later feed question generation and curation. In practice, it sits between raw course material and authored questions.

​
Core workflow

  1. Upload a PDF
  2. Wait for LearnTerms to process the document
  3. Review the generated chunks
  4. Edit, select, or remove chunks as needed
The downstream goal is not just storage. The point is to create high-quality source text that can be used in Question Studio and other content workflows.

​
Upload rules and limits

The current implementation is strict about uploads:
  • The documented library uploader is PDF-only
  • The configured PDF upload size limit is 16MB
  • Free users have a daily PDF upload limit of 1
  • Pro users effectively have a much higher daily ceiling
If the daily limit is reached, the app throws an explicit error. Free-tier messaging encourages the user to upgrade, while Pro-tier messaging treats the limit as a temporary daily cap.

​
What a chunk contains

Each chunk is intended to capture a meaningful section of the source material.
  • Title
  • Summary
  • Full content
  • Keywords
  • Type, such as paragraph, slide_group, diagram, table, or list
These fields matter because chunk quality affects generation quality. If the chunk title, summary, or content is vague, the resulting questions are more likely to be vague too.

​
Editing and curation

Admins and contributors can:
  • Rename documents
  • Update descriptions
  • Edit chunk titles, summaries, keywords, and content
  • Delete bad chunks
  • Select multiple chunks to support question-writing workflows
Document editing is also validated:
  • Title must be between 2 and 100 characters
  • Description is optional, but if present it must be between 10 and 500 characters
That validation is useful to document because it explains why some edits fail.

​
Scope and permissions

Content is cohort-scoped. A document belongs to a specific cohort, and the library query is filtered by that cohort ID. That means the library is not a global repository for every school or class in the system. It is intentionally local to the group that uses it.

​
Operational tips

  • Large PDFs may take a little time to process.
  • Clean chunk titles make later question-writing much easier.
  • Delete weak or duplicated chunks early so downstream generation stays focused.
  • Use descriptions to make document intent obvious when the title alone is not enough.

​
Common issues

  • No upload access: confirm the user has the right cohort role.
  • Unsupported file type: use PDFs for the documented upload flow.
  • Duplicate titles: document titles should remain unique inside a cohort.
  • Weak chunks: edit summaries and content before using them in Question Studio rather than trying to fix poor AI output later.