Buddy Icon Archive Preserving the tiny pictures that made the internet personal.

Conservation workbench

The machinery

Technical processing, enrichment, and moderation telemetry from behind the public collection census.

Live archiveRefreshes every 30 seconds

The collection

Tiny pictures, counted carefully

Every canonical icon represents one preserved image after known duplicate copies are grouped together.

Canonical icons preserved — the browsable collection
Original files stored — including grouped copies
Source collections — sites, packs, and recoveries
Source observations — recorded sightings with provenance
Animated icons — GIFs and other moving artifacts

Activity

What the archive has been doing

Loading the latest counts…

Archive collection, enrichment, and moderation activity in the last 5 minutes, hour, day, and 7 days
Window New icons Existing matches Match share Described Broad reviewed Narrow reviewed
Loading recent activity…
New iconsA file with a new exact fingerprint was validated and preserved.
Existing matchesA known file appeared somewhere else, adding source history without storing it twice.
DescribedAn icon received successful enrichment that makes it easier to search and understand.
Broad reviewedAn icon completed successful CPU moderation screening.
Narrow reviewedAn icon completed successful narrow moderation review.

Enrichment

Descriptions, schema by schema

Loading enrichment coverage…

Icons enriched — — of stored files have at least one successful record
Schema attachments — — successful records per enriched icon
Populated tag groups — — searchable tag assignments inside those groups
Extracted facts — text, subjects, colors, motion, identities, and other schema fields
Unique tag vocabulary — distinct searchable enrichment tags
Queue depth — Waiting for queue counts…
Current failures — — of completed attempts for the current metadata version
Average successful run — — total tokens per recorded response

Coverage

Still and moving icons

Static icons — / —
—
Animated icons — / —
—

Taxonomy

Most-used tag categories

  1. Loading tag categories…
Enrichment attempts grouped by prompt schema, metadata version, and outcome
Schema Metadata version Icons Stored coverage Successful records Failed records Success rate
Loading schema coverage…
Schema attachmentsSuccessful enrichment records. An icon may carry more than one schema or version.
Tag groupsDistinct tag categories populated within a successful record, such as subject, action, style, or color.
Extracted factsNormalized field values retained from the schema alongside searchable tags.
Queue depthQueued schema-and-icon combinations; first-pass and refresh work are counted separately.
Run averagesElapsed wall time and token totals across successful records that reported those measurements.

Content safety

Moderation, stage by stage

Loading moderation progress…

01

Broad screening

— / —

icons completed local classifier screening

Waiting for live queue counts…
02

Narrow review

— / —

icons completed detailed image and text review

Waiting for live queue counts…
03

Current labels

— / —

icons have a current moderation decision

Waiting for current label counts…
Broad screeningA pinned local model ranks risk and scheduling priority; it does not change what visitors can see.
Narrow reviewThe current image and extracted text receive the detailed policy assessment that can produce a site decision.
Current labelsStandard, review, and sensitive decisions are counted only for the active policy generation. Unknown icons remain public.

The work behind the gallery

Preservation depth

Finding a file is only the first step. The archive also records provenance, writes descriptions, and reconciles copies.

01

Descriptions

—

icons have searchable descriptions

— of stored files described ยท — remaining
02

Provenance

—

source-supplied tags retained

— tags per stored file on average
03

Duplicate grouping

—

known copies kept out of the main gallery

organized into — visual groups