AIPLA — AI in Physics Learning and Assessment

Execution repo for AIPLA, a technical-infrastructure contract for the University of Copenhagen Center for Digital Education: AI physics tutors, interactive sims, and teacher/researcher tooling for upper-secondary physics.

Repo: sunholo-data/cphu-aipla-app · Default & working branch: dev.

Built on the open-source AI Protocol Platform template (Skills + AG-UI + A2UI + MCP Apps + A2A on Google ADK). Some inherited files still say "Aitana Platform v6" — the AIPLA reality is authoritative. See CLAUDE.md.

Run it locally (no GCP credentials)

cp .env.example backend/.env      # then set LOCAL_MODE=1
cd backend && make install && cd ..
make dev                          # frontend :3456 · backend :1956

Open http://localhost:3456 — you get a yellow LOCAL_MODE banner and working demo skills, with Firestore/Firebase/Vertex all stubbed in-memory. Full walkthrough and the cloud tiers: WORKSHOP.md.

Where things are

frontend/	Next.js 15 + React 19 + Tailwind (chat, teacher, researcher surfaces)
backend/	FastAPI + Google ADK (skills, agents, protocols, auth, analytics)
cli/	the aiplatform CLI (make cli-install)
infrastructure/	mcp-sandbox (sim iframes) + Terraform
docs/design/	design docs — start at docs/design/README.md

Where the design lives

• Execution design (file paths, wire shapes, acceptance) → this repo, docs/design/aipla/SEQUENCE.md.
• Public project information → the app's native /project pages, authored as reviewed English Markdown in frontend/content/project/. Each page has a named owner, status, and next-review date; npm run check:project-content checks those fields and internal links before deployment.
• Historical product and pedagogical scoping material → the separate AIPLA Quarto repository. It remains an archive and redirect source during cutover, not the maintained public source of truth.
• Working with this repo as an AI agent → CLAUDE.md is the authoritative context (identity, gotchas, footguns, ports).

Common commands

make dev                              # local dev servers (:3456 / :1956)
make seed ENV=dev                     # push SKILL.md templates to Firestore (manual, post-deploy)
cd frontend && npm run quality:check  # CI parity: lint + typecheck + tests + build
cd backend && make lint && make test-fast

See the Makefile (make help) for the full target list and CLAUDE.md for the development workflow.