Skip to content
Contact support

Ask your course a question

For students

Alexandria answers from the materials your educator has placed in the course, meaning lecture slides, readings, and notes, rather than from the open internet. Each answer refers back to the page it came from, so it can be checked.

Two entry points, sharing the same interface:

  • The search bar in your library, for questions across everything you have access to
  • The chat tab inside a course, scoped to that course by default

Open Alexandria from your course in Canvas, Brightspace, or Moodle. Sign-in is automatic; no separate account or password is required.

The message “this course is not set up yet” indicates the educator has not yet connected the course.

Typing @ offers a list of the documents and folders you can scope to. Selecting one adds it as a scope, shown as a removable chip beside the question.

Scoping is worth the extra keystroke where a term carries different meanings in different weeks, or where a course is large enough that the relevant material is a small part of it.

Typing @ in the chat. Folders carry a folder icon, documents a page icon.

A selector on the chat sets how the answer is produced. Four modes are standard:

Mode Behaviour
Course Material Answers from your course materials, with citations. The usual choice.
General AI Answers from general knowledge, ignoring the materials. No citations.
Web Search Searches the web for answers, beyond the course.
Socratic (Study Mode) Does not supply the answer; guides you to it through dialogue.

A fifth kind of entry, a custom chatbot, appears only where your educator has configured one. A course with no chatbot shows the four above. Where several chatbots exist, each is listed separately under the name the educator gave it.

Your educator sets which mode you start in, and may have chosen Socratic deliberately.

The mode selector in a course with no custom chatbot configured, so four modes are offered.

Answers built from course material cite their sources. A line beneath the question states what the answer was built from and how many sources it used.

Selecting a citation opens the source: a page opens in the reader at the passage, a video at the timestamp, a web result in a new tab.

The sources line states what the answer was built from. The numbered chips are the citations for individual claims.

General AI answers state plainly that they were produced without the course material. Treat them as background rather than as course content.

Once an answer has finished, a row of three controls appears beneath it. The row is always visible, so there is nothing to hover for.

Control Function
Copy Copies the answer to your clipboard. The icon confirms briefly.
Regenerate Answers the same question again.
Question Generates a question on what you have just read.
The row beneath a finished answer: copy, regenerate, and Question.

Three tools are built into an answer and are easy to miss.

Select any text in an answer and a deepdive control appears. It opens an explanation of that specific idea in a side panel, without losing the conversation.

Deepdive on a selected sentence. The conversation stays in place on the left.

Test yourself. The Question control generates a multiple-choice question on what you have just read. The question appears beneath the answer with four options; selecting one marks it immediately and explains the reasoning against the text you were reading. A Try again control clears the selection so the same question can be attempted afresh. Useful for checking that an explanation actually landed. The question is not recorded anywhere: it is for you, not for your educator, and it does not count towards practice.

The Question control turns the answer just read into a multiple-choice question, marked as soon as an option is selected.

Interactive visualisations. Where a concept is better shown than described, an answer may include an interactive panel: a plot, a diagram, something you can manipulate. These appear on their own, and there is no control to request one. They are reserved for ideas that turn on a changing quantity, such as what happens to stability as you raise a time step, rather than anything a list of points would convey.

An interactive panel embedded in an answer. The controls recompute the figure as you change them.

The panel can be enlarged for a closer look using the control in its top right corner.

A PDF or plain text file of up to 10 MB can be attached to a question: your own notes, or a paper you found. Other formats are refused. A separate control attaches an image, under the same 10 MB limit. Attached files apply to that conversation only and are not added to the course.

Conversations are listed beside the chat and persist. They can be reopened, renamed, or deleted. New chat starts a fresh one.

Educators can see which topics generate questions across the course, in order to identify what is unclear while the course is still running. Conversation titles are visible to them; your name is not shown against them in that view.

This is aggregate teaching signal rather than assessment of individuals. Frequent questioning is the intended use of the tool.