Welcome to my personal portfolio repository. As a professional software developer, this site showcases my technical skills, key projects, and career achievements. It serves as a central hub for employers and collaborators to review my work and learn more about my background in software development.
The portfolio is structured for clarity and ease of navigation, allowing visitors to quickly find information about my experience, technologies I work with, and notable accomplishments. I keep this repository updated with my latest projects and milestones to reflect my ongoing growth in the field.
https://portfolio.anubhavlal.dev/ (also served at anubhavlal07.github.io)
Explore my portfolio for a comprehensive overview of my software development expertise, project highlights, and professional journey.
A static, dependency-free site: hand-written HTML, CSS and JavaScript with no framework, no third-party libraries and no build step. Content comes from Supabase (PostgreSQL) and is edited in the private portfolio-data dashboard. Local JSON snapshots are the offline fallback.
- HTML5 / CSS3: semantic markup, custom-property design tokens, Grid, Flexbox and container queries.
- Vanilla JavaScript (ES2020): one small renderer per section, with inline SVG icons.
- Supabase: content and visitor analytics, read through a tiny hand-rolled REST client (
assets/js/supabaseClient.js). - Google Fonts: Geist and Geist Mono.
The visual system is documented in docs/DESIGN.md.
anubhavlal07.github.io/
├── .github/workflows/snapshot.yml # Nightly Supabase → assets/json snapshot
├── assets/
│ ├── css/
│ │ ├── tokens.css # Colour, type, spacing, motion tokens (light + dark)
│ │ ├── base.css # Reset, typography, layout primitives, shared controls
│ │ └── nav, hero, work, experience, skills, about, resume .css
│ ├── img/ # Avatar, favicon, project screenshots, skill logos
│ ├── js/
│ │ ├── supabaseClient.js # Minimal Supabase REST client (global `supabase`)
│ │ ├── data.js # Loads all tables, falls back to JSON, normalises content
│ │ ├── icons.js # Inline SVG icons, maps ri-* names stored in the database
│ │ ├── nav.js # Header, scroll-spy, theme toggle
│ │ ├── hero.js # Status, headline, actions, featured project window
│ │ ├── motion.js # Cursor spotlight, tile glow, scroll reveal
│ │ ├── background.js # Interactive dot-field canvas (cursor glow, click ripples)
│ │ ├── work.js # Selected work bento and project flow diagrams
│ │ ├── experience.js # Experience list
│ │ ├── skills.js # Stack groups
│ │ ├── about.js # About, education and footer
│ │ ├── resume.js # Resume dialog (printable)
│ │ ├── analytics.js # Visitor analytics collector (sends to Supabase)
│ │ ├── disableInput.js # Blocks DevTools shortcuts, right-click and selection
│ │ └── pwa.js # Service worker registration
│ └── json/ # Raw snapshots of the public Supabase tables
├── docs/DESIGN.md # Design tokens, rules and the content data contract
├── projects/<slug>/index.html # Generated static page per visible project
├── scripts/snapshot.mjs # Writes assets/json from Supabase (no dependencies)
├── scripts/build-pages.mjs # Writes projects/ and sitemap.xml from assets/json
├── index.html # Single-page entry point
├── sitemap.xml # Generated: home plus every project page
├── sw.js # Service worker (network-first pages and JSON)
└── CNAME
assets/js/data.js requests the seven public tables (profile, social_links, skills,
skill_items, experience, projects, resume) from Supabase, and at the same time reads the
matching assets/json/*.json snapshots:
- The snapshot renders first, so the page is never empty.
- When Supabase answers, every section re-renders with live data. Any table that fails falls back to its snapshot.
- Renderers subscribe with
Site.onContent(fn)and receive a normalised content object (seedocs/DESIGN.md).
Everything visible on the page is editable from the dashboard, including the hero headline
(profile.tagline, falling back to your name), the focus chips (profile.works_on) and each
project's flow diagram (projects.flow, falling back to the tech list in subtitle). Visibility,
featured state and ordering follow is_visible, is_featured and display_order.
scripts/snapshot.mjs writes the snapshots using the public anon key, and
.github/workflows/snapshot.yml runs it every night (and on demand from the Actions tab),
then runs scripts/build-pages.mjs, committing assets/json, projects/ and sitemap.xml only
when something changed. To refresh by hand: node scripts/snapshot.mjs && node scripts/build-pages.mjs.
Every visible project also gets its own crawlable page at /projects/<slug>/, written as plain
static HTML by scripts/build-pages.mjs (Node 22, no dependencies). The generator reads
assets/json/projects.json, profile.json, social_links.json and resume.json, and reuses the
site's own data.js (normalisation and slugs), work.js (flow diagram markup) and icons.js, so a
project page renders exactly what the home page tile shows.
- Slug: the title lowercased, every run of characters outside
a-z0-9replaced with-, and leading/trailing-trimmed (Document Automation Platform→document-automation-platform). The dashboard database uses the same rule (_project_slug), so clicks and pages line up. - Content in the HTML: title, description, tech chips (from
subtitle), the "How it works" flow whenprojects.flowis set, the screenshot, the primary link, other projects and the footer. - Head: a unique
<title>(<Title> — Anubhav Lal), a meta description of at most 160 characters (built from the title and tech when the description is empty), the canonical URL, Open Graph and Twitter tags using the project image, and JSON-LD (SoftwareSourceCodewithcodeRepositoryfor GitHub links,CreativeWorkwithurlotherwise, plus a breadcrumb). - Design: the pages load the same
tokens,base,nav,work,aboutandresumestylesheets with root-relative paths, the same theme bootstrap and theme toggle, and the resume dialog. - Housekeeping: pages for projects that are hidden or deleted are removed, and
sitemap.xmlis rewritten with the home page and every project page (lastmodfromupdated_at). Unchanged inputs produce byte-identical output; the footer year comes from the build date (BUILD_YEARoverrides it).
The home page links each tile to its page with a Details link.
Add ?ref=<tag> to any link you share and the dashboard groups visits by it (first touch per
session). utm_source and utm_campaign work too when ref is absent. Tags are lowercased and
kept to a-z 0-9 . _ / -, at most 64 characters.
| Where the link lives | Link |
|---|---|
| LinkedIn featured section | https://portfolio.anubhavlal.dev/?ref=linkedin-featured |
| Resume PDF | https://portfolio.anubhavlal.dev/?ref=resume |
| A GitHub README | https://portfolio.anubhavlal.dev/projects/cerebro/?ref=github-cerebro |
| A newsletter | https://portfolio.anubhavlal.dev/?utm_source=newsletter&utm_campaign=october |
The tracker sends the full page_url, query string included, so nothing extra is needed on the site.
- In Google Search Console add a Domain property
for
anubhavlal.dev. - Copy the
google-site-verification=…TXT record it shows, add it at the DNS provider foranubhavlal.dev, wait for it to propagate and press Verify. - Open Sitemaps and submit
https://portfolio.anubhavlal.dev/sitemap.xml. - Use URL inspection on a project page (for example
/projects/cerebro/) and request indexing to speed up the first crawl.
- Responsive: mobile-first layouts from 320px to wide desktop; the four nav links stay visible on every screen size.
- Theme: auto, light and dark, applied before first paint and persisted in
localStorage. - Project flows: each project's pipeline drawn from dashboard data, featured in a framed hero window.
- Resume dialog: native
<dialog>with download and print. - Accessibility: skip link, visible focus, labelled icon links, reduced-motion support.
- Offline: installable PWA with a network-first service worker.
- Visitor analytics: anonymous session, device and engagement metrics sent to Supabase.
assets/js/analytics.js records a visit only after the page has been open for 5 seconds, and
?notrack opts a browser out (?track opts back in). Once the visit is recorded it also sends:
- Section views: each
main section[id]andfooter#contactcounts once per page load when it is at least 40% visible, or fills at least half the viewport, for one second. Each one callsrpc/record_section_viewwith{ p_session_id, p_section }. Home page ids arehome,experience,work,skills,aboutandcontact; project pages useproject,flow,moreandcontact. - Link clicks to
link_clicks, withclick_kindandproject_slugwhen the page knows them: the nav Resume link isresume_open, the dialog's "Download PDF" isresume_download, links carryingdata-project="<slug>"areproject_code(GitHub) orproject_demo,mailto:links areemail, and LinkedIn, Instagram, X/Twitter and Medium links aresocial. Anything else, including GitHub links withoutdata-project, is left for the database trigger to classify. Same-site links (Details, back links) are not recorded as clicks.
Serve the folder over HTTP (opening index.html directly with file:// breaks the fetch()
calls):
python -m http.server 8000
# then open http://localhost:8000Local visits are recorded by analytics.js like any other visit.
Hosted on GitHub Pages at the custom domain in CNAME. There is no build step: pushing to
main publishes the site. When shell files change, bump CACHE in sw.js so returning visitors
get the new version. After editing projects in the dashboard, run node scripts/build-pages.mjs
(or wait for the nightly workflow) so the project pages and sitemap follow.