AIPLA — University of Copenhagen AIPLA — University of Copenhagen
  • Home
  • About
  • Strands
  • Examples
  • Timeline
  • Architecture
  • Evaluation
  • Self-hosting

KineBot v2

Hybrid AI kinematics tutor — NCERT/CBSE Class 11 physics, built by DK

KineBot v2 is a feature-complete AI tutoring system for Class 11 kinematics. It represents a distinct artefact class from AIPLA’s physics simulations: a hybrid AI platform combining simulations, Socratic chat, adaptive quizzes, and reference material in a single self-contained file.

Built by DK for an Indian student cohort. DK’s students (~100s available) are the beta audience for AIPLA’s first non-Danish curriculum track.

What’s inside:

  • 7 canvas-based simulations — 1D uniform motion, uniformly accelerated motion, free fall, projectile motion, circular motion, vector addition, relative velocity; each with parameter sliders and live position/velocity/acceleration readouts
  • Motion graph plotter — x-t, v-t, a-t curves; range-vs-angle and max-height plots for projectile motion
  • Socratic chat tutor — Claude, warm and encouraging persona, covers all 11 kinematics topics, quick-prompt buttons (requires API key — see note below)
  • Adaptive quiz — AI-generated MCQs filtered by current topic (requires API key)
  • Formula reference — static lookup by topic

Chat and quiz require an Anthropic API key entered in the browser. The seven simulations and graph plotter work without a key. This is the as-shipped form; AIPLA’s migration replaces the browser-side key with central key management (ADR-014) so students never handle API credentials. See the migration brief for the full integration spec.

Open in a new tab to use chat and quiz with your own API key.

AIPLA migration summary (skill 2):

The migration converts KineBot from a browser-side-key artefact to an AIPLA-compliant activity. Three changes to the source file; everything else stays:

  1. Remove three direct fetch() calls to api.anthropic.com and the API key input UI
  2. Route chat and quiz through AIPLA’s backend via postMessage
  3. Add state events (topic change, simulation run, graph switch) so the tutor skill reads what the student is doing

This migration doubles as a reusable intake pattern — the checklist it produces is the standard process for onboarding any externally-built AI artefact into AIPLA.

← Back to Examples

AIPLA — Center for Digital Education, University of Copenhagen

 
  • Edit this page
  • Report an issue

Built with Quarto