Get started¶

Open the Gradebird application and use the sign-in method configured for the deployment. Standard deployments use email/password accounts and require email verification. Some institutional deployments route anonymous users directly to organization single sign-on and may hide local account-management links.

After verification, the account belongs to one organization. Access may be immediate, invitation-based, or subject to approval. A user awaiting approval sees a pending-approval page.

Create a course¶

From the app home, select Create course. The initial form asks for Course name, Term, and Institution. It does not require an LMS choice. After saving, the course owner enters guided setup.

Accounts that are graders only cannot see or open course creation. An invited instructor can work in the shared course after accepting the invitation; the owner's setup progress is not imposed on other instructors.

Guided setup¶

Guided setup is owner-only and begins after creating a new course. Existing and copied courses do not enter it.

The setup menu contains four independently skippable steps:

Course context — enter course-level background used by collaborators and optional AI workflows.
Agreement tolerance — set the Agreement tolerance (points), the allowed difference between agents' scores in points. Eligible results agreed by at least two agents are automatically saved as Provisional, awaiting human approval. Choose Queue review for individual review and candidate comparison, or Bulk review for any status and inline saved-assessment edits. The page explains how model agreement determines whether you review one assessment or compare differing assessments. The initial tolerance is 1 point. Grading mode and the full agent configuration remain available later under AI Settings → Advanced settings; we recommend keeping the defaults.
Roster — optionally upload a Canvas or D2L CSV. A valid preview can detect and save the LMS format when the course has none.
First exam — create an exam and optionally launch the exam-header tour.

Save and continue advances from the current step. Skip skips only that step. Skip all remaining setup finishes setup and opens the explicit Exams tab. Returning to the course root resumes the next unfinished step until setup is finished; selecting Exams directly remains available.

For Gradebird-generated packets, Start the exam builder tour is checked by default during course setup, including later courses. It is not offered for your own print packets. The tour starts with importing an existing exam and explains builder setup, the exam label, version selector, publish control, and contextual help; it does not alter exam content.

Resume ordinary work¶

Use the course-name menu in the top bar to move between Exams, Roster, AI Settings, Access, Gradebook, and Analysis. Edit the course name, term, institution, or LMS setting through the small edit control beside the course heading.

Before printing, create or import an exam, compile a successful preview, publish a version, and wait for its packet PDF to become available. A roster is optional for authoring and for master or numbered packet printing, but student matching, named packets, email delivery, and LMS export depend on roster data.

Create course¶

This page creates a new course workspace. Enter the instructor-facing course details and submit the form; a successful creation opens guided course setup. Graders cannot create courses, and no course is created while the form contains errors.

Course setup: Course context¶

This first setup page records optional course context for later AI-assisted workflows. Save the context to continue, or use the setup navigation to revisit an unfinished step. Saved changes affect newly launched AI work rather than already queued or completed work.

Course setup: Agreement tolerance¶

This setup page sets the default agreement tolerance in points. Agreement between at least two models automatically creates eligible provisional assessments for human review; there is no separate switch to enable this. Bulk review remains available for every assessment status. The settings can be changed later from AI settings, and each question can override the default from its Assessment tab. After a policy change, that tab can reconcile eligible AI assessments into Provisional or Needs review without changing their original AI run records.

Course setup: Import roster¶

This optional setup page imports a Canvas or D2L Brightspace roster and presents the same review-before-apply workflow as the regular roster page. You may continue without importing; skipping the step does not create or deactivate students.

Course setup: Create first exam¶

This final setup page creates the course’s first exam workspace or opens an existing exam. The optional tour opens after navigation to the builder. Completing or skipping this step ends guided setup and returns the course to its normal work areas.