Education division of Sand Point Studios Updated September 2026
Agoge

Assessment and course tooling for college teaching

Agoge CourseOps · in pilot with faculty

The whole stack, on one instructor's machine

A course development studio where a course is a folder of files, edited in conversation with an AI assistant, previewed as it will look in the LMS, and exported as a Brightspace import package, with the graders alongside it.

§ 1

The loop it is built around

The CourseOps iteration loop Four stages inside a dashed boundary labelled your machine: draft, generate, review, publish. A revise arrow runs from review back to draft. Only the Brightspace package crosses the boundary, plus one optional call to the AI assistant you signed in to. YOUR MACHINE The course is a folder of files. Nothing here leaves it. 01 Draft outline, syllabus 02 Generate assignment, rubric, key 03 Review grade a test submission 04 Publish locked D2L package revise, as many times as it takes To the LMS optional: your own AI assistant, on your own account
Figure 2. Everything but the last arrow happens on your machine. The course material leaves only when you publish it, and the one dotted line is the assistant you chose, billed to your own account.

§ 2

Bring your own model

CourseOps has no AI account of its own to sell you. Every generator goes through a single bridge to whichever assistant you signed in to: a ChatGPT Edu or Plus seat your institution already pays for, or an API key you hold.

The cost lands on an account you control, and the question of which vendor holds your course material is one you answer rather than one we answer for you.

§ 3

Much of it needs no model

Seven of the eight generators produce something real with no assistant configured: a full dataset, a rendered deck, a starter workbook, a Power BI project, a Tableau workbook.

The graders split the same way, and the interface shows which is which before you run anything. Whether a cell holds a formula is settled by inspection; whether it is the right chart for the data is a judgement that needs a model.

§ 4

Where the course material sits

At the local tier this stops being a policy promise and becomes a fact about the network. The container writes to a folder in your Documents directory. On a managed university machine that folder is usually already synced by OneDrive, so backup is something you have rather than something you configure. Moving to a new machine is copying a folder, and courses move between installs as a zip, which is also how a course gets handed to a colleague.

§ 5

The same application at three sizes

TierWhoAddsNeeds
Local appOne instructor, own machine, own AI sign-inCourses shared by zip export and importNo IT involvement
Department serverA college or departmentCo-instructors, graduate and undergraduate assistants on shared coursesA server and backups
Campus PlannedThe whole institutionSingle sign-on and direct publishing into BrightspaceA real deployment

Where CourseOps is today: it is the newest thing we have built, in pilot with faculty rather than generally available, and the campus tier is design work rather than running code. The local tier is real. It installs from a public container image, activates against a licence and updates itself in place.

Ask to see the running application