Syrka landing page

The syrka.co homepage: a single Next.js page (app/page.tsx) built from the components in components/landing/. Campus and sign-in links point to https://syrka.co.

npm install
npm run dev     # http://localhost:3000
npm run build

Background videos live in public/media/ (960×540 H.264, 10 s, no audio). Bump the ?v= query on a video's URL in app/page.tsx after re-encoding it in place, or browsers that cached the old file will stall.

SYRKA Campus demo (/demo)

An interactive demo of SYRKA Campus for final-year students and employers. The demo persona's profile, projects and assessments are fictional and labelled as such; external modules show real, attributed data. Rule: external data describes the job market, open work or national context — it never becomes a student's skill, verification, project evidence, assessment result, capability score or activity.

What is where

• Routes: /demo (choose a persona), /demo/student/* (dashboard, profile, skills, projects, opportunities, open-source missions, visibility, data sources), /demo/employer/*, /demo/admin.
• Personal data: lib/campus/store.ts keeps each visitor's copy in localStorage (profile edits, assessments, projects, saved and dismissed opportunities, saved missions, connected repositories, activity). Schema changes go through lib/campus/migrations.ts, which upgrades a visitor's saved state instead of discarding it. Reset demo data restores the seed. The demo student starts with no activity; every event is something the visitor did.
• Matching: lib/campus/matching.ts, one engine for both portals (formula documented in the file).
• External sources: registered in lib/campus/sources.ts (use, licence, attribution, refresh, limits) and listed at /demo/student/data-sources.

Module	Source	Code	Refresh
Real job openings	Company careers pages (Greenhouse, Lever, Ashby feeds); Himalayas optional	lib/campus/live-jobs.ts, live-jobs.server.ts	Cached ≤12 h; Apply goes straight to the company's posting
Live market demand, Skills in demand now	Computed from live openings	lib/campus/market-demand.ts	With live openings; daily snapshot for 7-day change
Missing-skill relevance	O*NET 31.0 Software Skills (internal mapping only)	lib/campus/onet-role-skills.ts (generated)	Static; npm run campus:onet
Open-source missions, connected repositories	GitHub REST API (read-only)	lib/campus/github.ts	Missions cached 6 h with ETags; repositories on connect/refresh
Explore related open work	OpenAlex	lib/campus/open-data.ts, app/api/campus/open-work	Cached 24 h per query
India career landscape	World Bank WDI (incl. ILO modelled estimates)	lib/campus/open-data.ts	Cached 24 h; series older than 4 years hidden

Environment variables (all optional)

Variable	Purpose
CRON_SECRET	Protects /api/campus/refresh. Vercel Cron sends it automatically. Without it the route refuses every call.
CAMPUS_HIMALAYAS	on adds remote roles from Himalayas (credited, applied through Himalayas). Off by default.
HIMALAYAS_API_KEY	Raises Himalayas rate limits when it is on (server-side only).
CAMPUS_LIVE_JOBS	off hides real openings and live demand (e.g. investor environment); illustrative roles are shown, clearly marked.
CAMPUS_LIVE_MAX_AGE_HOURS	Hide live data older than this (default 36).
CAMPUS_COMPANY_BOARDS	Companies whose careers pages are read, e.g. greenhouse:groww:Groww,lever:meesho:Meesho. Defaults to the 20 companies in DEFAULT_COMPANY_BOARDS (Razorpay, Paytm, Meesho, Groww, Stripe, MongoDB…).
GITHUB_TOKEN	A fine-grained token with no permissions (public read only). Raises GitHub's limit from 60 to 5,000 requests/hour.
CAMPUS_GITHUB_REPOS	Overrides the mission allow-list (below).
OPENALEX_API_KEY	Free key from openalex.org/settings/api. Without it, requests share a very small daily allowance per IP.
KV_REST_API_URL + KV_REST_API_TOKEN (or UPSTASH_REDIS_REST_URL + UPSTASH_REDIS_REST_TOKEN)	Stores daily market-demand snapshots (aggregate counts only) for the 7-day change. Without it, production shows current demand only; local development writes to .data/.

Open-source mission allow-list

Missions are read only from repositories in the allow-list — never from search. The initial list (DEFAULT_MISSION_REPOS in lib/campus/github.ts, checked 2026-10-26 for open beginner issues):

freeCodeCamp/freeCodeCamp (first timers only, help wanted), scikit-learn/scikit-learn, jupyterlab/jupyterlab, kubernetes/website, microsoft/vscode, vercel/next.js, oppia/oppia, excalidraw/excalidraw (good first issue), processing/p5.js (Good First Issue), python/mypy (good-first-issue).

To change it, set CAMPUS_GITHUB_REPOS to comma-separated owner/repo:label one|label two entries (labels default to good first issue; each label is queried separately). Add only reputable public repositories. Pull requests, assigned and locked issues are excluded automatically.

Scheduled refresh

GET /api/campus/refresh (requires Authorization: Bearer $CRON_SECRET) re-reads the live sources, stores today's demand snapshot and marks cached live jobs, missions and landscape data for background refresh. On Vercel, vercel.json schedules it daily at 01:30 UTC (the Hobby plan allows daily crons). Pages also refresh their caches on their own when visited; the cron makes sure a snapshot exists every day. Run it by hand with:

curl -H "Authorization: Bearer $CRON_SECRET" http://localhost:3000/api/campus/refresh

Local development

npm install
npm run dev            # http://localhost:3000/demo
npm run test:campus    # matching, live sources, external modules, honesty rules
npm run campus:calibrate   # re-derive demo job requirements from live postings
npm run campus:onet -- <folder with O*NET software_skills.csv>   # rebuild the O*NET mapping

All external modules work without keys at low volume. Each shows an empty state, never sample data, when its source is unavailable.

Calibration of illustrative roles

npm run campus:calibrate reads entry- and mid-level postings and regenerates lib/campus/calibration.ts (aggregate skill shares only). Demo job requirements are derived from it in lib/campus/seed.ts (≥50% of postings → required, ≥30% → preferred, at least 15 postings and 2 required skills, otherwise hand-written). Bump SCHEMA_VERSION and add a migration step after recalibrating.

Attribution

Includes information from the ONET 31.0 Database by the U.S. Department of Labor, Employment and Training Administration, used under CC BY 4.0. Syrka has modified and mapped portions of this data. USDOL/ETA has not approved, endorsed, or tested these modifications. ONET® is a trademark of USDOL/ETA. Job openings come from company careers pages (and Himalayas when enabled); missions and repository metadata from GitHub; research metadata from OpenAlex (CC0); India indicators from the World Bank (CC BY 4.0).