The future of learning, available today.
Get cited answers from the instructor's own course material, inside Canvas.
Example of a Pensa answer: a student asks how A* search compares to Dijkstra's algorithm. Pensa answers from the course material, and every sentence carries a citation to the exact file and page, for example Lecture-06.pdf, page 12. The answer is marked Grounded.
ALG2100 · CanvasHow does A* search compare to Dijkstra?
Dijkstra's algorithm always expands the node closest to the start, exploring evenly in every direction. Lecture-06.pdf· p. 12 A* adds a heuristic estimate of the remaining distance, so the search leans toward the goal and visits far fewer nodes. Lecture-06.pdf· p. 14 With an admissible heuristic, A* still finds the optimal path. That is exactly what exercise 2a asks you to show. Exercise-set-6.pdf· p. 2
All AI processing stays in the EU
Backed by and developed with
Students already use AI. Institutions have lost track.
Course material gets uploaded to chatbots nobody at the institution controls. The answers skip the course, and the learning happens out of sight.
Course material leaks out
Notes and exercises end up in tools with unknown data practices.
Answers detached from the course
General chatbots answer from the internet, not from what the course teaches.
One pace fits fewer than ever
Fixed lectures shut out students with caring duties, dyslexia, or a job on the side.
9 in 10
students already use AI in their studies.
Studiebarometeret 2025, HK-dir
The lecture hall hasn't changed. The students have.
Reach more students.
Students will use AI either way. Pensa is the version made for the learning institution.
Control comes back
Course material stays under the institution's agreement. The instructor decides what Pensa reads.
For the institution
One governed tool replaces a hundred private ones.
Learning stays on the curriculum
Every answer comes from the instructor's own files, cited to file and page.
For the instructor
The AI teaches your course, in your definitions.
More students keep up
Students adapt Pensa to their needs. You see what they wonder about most, even the things they never ask out loud.
For students
The subject meets the student, not the other way around.
Developed with instructors at NTNU.
No new tool. No new login.
Pensa is a tab inside the Canvas course students already open.
Privacy is a core focus.
Pensa was built around protecting student and institutional data from day one.
All AI processing stays in the EU
Every model call runs on EU infrastructure.
No tracking cookies
Sessions are kept server-side, under the institution's agreement.
Chat history is deleted automatically
Anything older than six months disappears on its own.
Students can delete everything, anytime
One action, and the history is gone.
The instructor sets the boundaries
What Pensa reads and how it helps: the instructor decides.
Built at NTNU, with instructors.
Pensa is built at NTNU, with instructors and the AI community.
Jakob Ildstad
Founder · Engineering
Cybernetics & Robotics, NTNU
Theo Stene
Co-founder · Engineering
Cybernetics & Robotics, NTNU
William Dischler Hjelkrem
Co-founder · Business
Economics, NTNU
See Pensa in your course.
We'll set up a demo with your own course material.