zerodig

Advanced YouTube Search. A client-side, single-page interface to the YouTube Data API v3 that exposes every filter the API offers in one place. Search videos, channels and playlists; narrow by date, region, language, duration, definition, license, category and more; then refine locally by views, likes, comments and length. Black/white terminal aesthetic. No server, no build step, no tracking.

▶ Live site: https://flazeiguess.github.io/zerodig/

Open it, paste a free YouTube Data API key, and search. Everything runs in your browser; nothing is stored on a server.

---

What you can filter by

Native search.list parameters (server-side):

Filter	Options
query q	free text with "exact phrase", `a
type	video · channel · playlist · all
order	relevance · date · rating · view count · title · video count
published after / before	any date window
region · language	ISO country + language codes
safe search	moderate · none · strict
within channel	restrict to a channel id
duration	short (<4m) · medium (4-20m) · long (>20m)
definition	HD · SD
dimension	2D · 3D
captions	has captions · none
license	standard YouTube · Creative Commons
video type	movie · episode
category	Music, Gaming, Education, News …
event	live · upcoming · completed
embeddable · syndicated · paid promotion	yes / any
location + radius	geo search around a lat,lng
results per page	1-50

Local post-filters (client-side, what the API can't do):

• min / max views
• min likes
• min / max length (seconds)
• re-sort results by views ↑/↓, likes, newest, oldest, longest, shortest

How it works

The API can't filter or sort by view count, likes or exact length, so the app runs a two-stage workflow, client-side:

1. search.list with your native filters (100 quota units / page, up to 50 ids).
2. Enrich each page depending on result type: videos.list (statistics + contentDetails), channels.list (subscribers, video count) or playlists.list (item count), 1 unit each.
3. Filter & sort the enriched results locally by views, likes and length.
4. Auto-page: follow nextPageToken until it hits your dig depth or stop-after target. Load more continues from where it stopped.

The live console prints every page fetched and the running quota cost.

Quick start

1. Open the live site.
2. Get a free API key (below) and hit Save.
3. Type a query, pick a type, tune filters, and hit Search.

Get an API key (free)

1. Go to the Google Cloud Console.
2. Create a project, then APIs & Services -> Library -> enable YouTube Data API v3.
3. Credentials -> Create credentials -> API key.
4. Paste it into the app and hit Save (stored in your browser's localStorage).

Key security

Because this is a static site, the key sits in your browser and travels only to Google, but anyone who opens the deployed page and views the network tab could read it while you use it. Protect it:

• Application restriction -> HTTP referrers -> add the Pages domain: https://flazeiguess.github.io/zerodig/*
• API restriction -> restrict the key to YouTube Data API v3 only.
• Daily quota is 10,000 units/day. A search page is 100 units + up to 3 to enrich; a depth-3 run is roughly 309 units. The status bar tracks today's usage.

If you ever want the key fully hidden, move the fetch calls behind a tiny proxy (Cloudflare Worker / Django) that holds the key server-side. The frontend logic stays identical.

Run locally

Plain HTML/CSS/JS. Any static server works:

python -m http.server 8000
# or
npx serve .

Then open http://localhost:8000.

Deploy (GitHub Pages)

Already published to Pages from main at https://flazeiguess.github.io/zerodig/. To reproduce on a fork:

1. Push the files to a repo.
2. Settings -> Pages -> Source: Deploy from a branch -> main / root.
3. The included .nojekyll file serves everything as-is.
4. Add your own Pages domain to the API key's referrer restriction (above).

Assets are versioned with a ?v= query so a redeploy busts the browser cache. Bump it (e.g. app.js?v=3) in index.html when you ship changes.

Files

File	Role
index.html	structure
tokens.css	design tokens (colour, type, space, motion)
style.css	terminal styling
app.js	search, enrichment, filtering, paging
.nojekyll	serve raw on GitHub Pages

Notes

• Video-only attributes (duration, definition, category …) apply to type = video and hide for channel / playlist searches.
• View / like / length filters apply to video results; channels and playlists pass through unfiltered.
• Videos with hidden statistics show n/a and are excluded when a numeric view filter is active (they can't be verified).
• Everything is grayscale on purpose except the thumbnails, which show in full colour.