This is the engine yt0 behind the artwork Arranging Data Into Lists.
Software that facilitates the downloading of video content that has not been viewed yet. Exhibited as a video installation that displays a montage of the dataset on an inaccessible screen.
Live: https://v0.cnrd.computer/ (original gallery at /v1)
- Crawls public YouTube search results and channel feeds. It uses no API key and no login.
- Archives videos that show zero views and are at least seven days old.
- Verified: the zero is confirmed by the channel feed.
- Unverified: the video only looked like zero once.
- Paces itself to at most one request per second and one search per minute, and backs off when YouTube pushes back.
- Serves the archive through a small read-only API and a browser interface. Clustering and search run entirely in the visitor's browser.
| Path | What |
|---|---|
main.ts |
Starts the crawler and the web server |
crawler.ts, provider.ts, network.ts |
Discovery, verification, request pacing |
store.ts |
SQLite archive and task queue |
server.ts, public/ |
API and web pages |
cli.ts |
Seed, import and backup commands |
deploy/ |
systemd units and proxy route |