UCSD's classic WebReg, brought back as a class-planning tool — for everyone who misses it now that TSS is the only option. It looks and works like the WebReg you know, loaded with real Fall 2026 courses, professors, times, rooms, and total seat counts.
Just open the link. It runs entirely in your browser, and your planned schedule is saved privately on your own device.
Two hard problems: reproduce WebReg's exact interface, and get Fall 2026 data that only lives inside UCSD's new SAP system.
WebReg was already retired, so there was no live page to copy. I rebuilt it from primary sources — the UCSD Registrar's own "WebReg 2.0" demo deck and the colleges' enrollment guides — reconstructing the layout screen by screen and sampling the exact colors out of the screenshots. It's plain HTML, CSS, and vanilla JavaScript (no framework) to match the old ExtJS-era feel and stay dependency-free: the collapsible search drawers, the 13-column section grid, the enroll/plan dialogs, and the List / Calendar / Finals schedule views are all rendered client-side, including a weekly calendar grid that lays out overlapping blocks and flags time conflicts.
The legacy public Schedule of Classes only goes through Summer 2026. Fall 2026 exists only in TSS — UCSD's new system — which is SAP Student Lifecycle Management behind an SAP Fiori launchpad, gated by SSO + Duo. And its own class-search screen is unreliable and crashes.
I wrote a Playwright script that opens a real Chromium window at the TSS login, waits for me to finish SSO + Duo, and captures the authenticated session. (A subtle bug: login detection had to key on a real SAP_SESSIONID cookie — the Fiori URL loads before authentication and kept false-positiving, closing the window mid-login.)
Since the search UI itself crashed, I didn't try to drive it. Instead I booted the Schedule-of-Classes Fiori app headless with the captured session and watched its network traffic. That exposed the real data source the launchpad never advertises: an OData v4 service, yucsd_con_module. I queried it directly — found the term value-help (Fall 2026 = year 2026, period 2), then paged the entity sets and dumped the whole term to disk: 1,768 courses, 8,431 section-events, 7,408 meeting times, and 7,107 instructor records — bypassing the broken interface entirely.
SAP models a course as a "module" and its sections as "events" bundled into enrollment "packages" (a lecture with its discussion and lab). I wrote an importer that regroups these into WebReg's structure — lecture groups (A / B / C…), discussion rows (A01…), lab rows (A50…), and a final-exam row — deduping meetings that are shared across packages and parsing SAP's human-readable schedule strings like
Tu, Th 08:00 AM - 09:20 AM In Person @ Center Hall Room 105
Final Examination 12/08/2026 08:00 AM - 10:59 AM In Person
into structured days, times, rooms, and final-exam entries.
For development this is a Flask + SQLite app with a small JSON API. To hand it to non-technical friends as just a link, I converted it into a fully static site: a build step bakes the entire catalog into one JSON file, and a small shim (site/js/localdb.js) overrides window.fetch for the /api/* routes so the exact same frontend runs entirely in the browser — client-side search, and each person's schedule stored privately in localStorage. No server, no database, no shared state, nothing uploaded anywhere.
It's hosted on GitHub Pages. A GitHub Actions workflow rebuilds the static site from source and deploys it on every push to main (scripts/build_site.py: seed the DB from committed course data → bake the catalog → assemble site/), so any code change or data refresh goes live in about a minute.
# Local dev (full Flask app)
pip3 install -r requirements.txt
python3 seed.py # builds data/webreg.db from the committed course data
python3 app.py # http://localhost:5070 (5060 is browser-blocked as a SIP port)
# Build & preview the static site
python3 scripts/build_site.py
cd site && python3 -m http.server 5090 # http://localhost:5090Refreshing the Fall 2026 snapshot (seats/waitlists are captured at build time):
python3 tss/connect.py # opens a browser; log in with UCSD SSO + Duo
python3 tss/import_fa26.py # map the captured TSS data into WebReg's format
python3 scripts/build_site.py # rebuild; commit + push to main to auto-deploySession cookies live in tss/state.json, which is git-ignored and never leaves your machine.
site/ the deployed static site (index.html, css, js, data, localdb.js)
scripts/ build_site.py, export_static.py — build the static bundle
app.py Flask server + JSON API (local dev)
seed.py build data/webreg.db from data/parsed/
schema.sql SQLite schema
data/parsed/ committed course data (per term, per subject)
scraper/ legacy Schedule of Classes scraper
tss/ TSS SSO capture + FA26 importer
templates/ static/ the WebReg UI (source of the static build)
docs/ DESIGN.md, UI spec, research notes, screenshots
.github/workflows/ auto-deploy to GitHub Pages on push
Released under the MIT License — free to use, modify, and share.
Not affiliated with or endorsed by UC San Diego. Independent, non-commercial planning tool — it doesn't perform enrollment or access your UCSD account; data is read from UCSD's own systems with your own login and stays on your device. Always confirm details in TSS before enrolling.