Canvas is the beating heart of day‑to‑day teaching at many universities, so ibl.ai was built to feel like a native part of the LMS rather than an add‑on. The integration hinges on LTI 1.3 Advantage, the latest version of the IMS Global standard that Canvas fully supports. Once the tool is installed, every user—whether student, instructor, or TA—arrives already authenticated, already enrolled, and already inside the right course context. That removes the two biggest adoption hurdles for generative‑AI tools: extra logins and manual roster setup.
From a technical perspective, Canvas supplies a secure, signed OIDC launch token that carries the user’s Canvas ID, role, course, and LTI service URLs. ibl.ai validates that token, spins up an in‑page session, and can immediately call Canvas APIs with the delegated scopes that the admin granted during installation. Because LTI 1.3 includes Advantage services such as Names & Roles and Assignments & Grades, the same launch also hands ibl.ai endpoints for roster sync and grade passback—no extra API key dance required.
Core Integration Features
LTI 1.3 Launch – Canvas sends a signed OIDC JWT that logs the user in and passes course context; ibl.ai loads inside an iframe—zero extra clicks.
Roster Sync (NRPS) – Names & Roles service or the /enrollments API keeps ibl.ai’s user list fresh for analytics and personalization.
Deep Linking – Instructors insert ibl.ai content or assignments via the External‑Tool picker; ibl.ai returns link metadata that Canvas renders like any native item.
Gradebook Passback – AI‑generated scores and formative comments flow back to SpeedGrader through the LTI Score service or Submissions API.
OAuth2 Scopes – Least‑privilege Developer‑Key scopes (roster, grading, files) fence what ibl.ai can touch; everything is audit‑logged in Canvas.
What a Typical Session Looks Like
1. Launch – A student clicks the "Weekly Writing Coach" assignment created via deep link; Canvas issues an LTI token and drops ibl.ai into the page.
2. AI Session – ibl.ai reads rubric criteria from the launch, fetches the student’s draft from Canvas Files, and delivers inline feedback while the student edits.
3. Roster & Context Checks – If someone just added the course, ibl.ai calls NRPS to pull the latest roster so analytics stay accurate.
4. Passback – When the student submits, ibl.ai posts a rubric‑aligned score and feedback comment; SpeedGrader updates instantly and the student sees results in their gradebook.
Faculty & Student Value
The tight Canvas integration means instructors don’t change their workflow: they still create Assignments, set points, and view SpeedGrader. The only difference is that a powerful AI mentor sits behind the "External Tool" link doing the heavy lifting—draft feedback, hint generation, or automated rubric scoring. Students benefit because they never leave the familiar Canvas interface and never have to juggle accounts or copy‑paste work. Everything—AI chat, file uploads, feedback, and grades—lives in one place and one audit trail.
Administrators, meanwhile, gain single‑pane visibility: ibl.ai usage, grades, and engagement metrics all flow into Canvas analytics dashboards. Because LTI scopes are explicit, IT can satisfy privacy officers that ibl.ai can’t reach other courses or institutional data. And if policies change, disabling the tool is as simple as flipping a Developer‑Key toggle.
In short, by speaking Canvas’s native LTI dialect and REST APIs, ibl.ai becomes a first‑class citizen of the LMS—rolling generative AI into existing teaching practices without extra passwords, CSV imports, or policy exceptions.
Learn more at ibl.ai