Skip to main content
Glama

Pelican Bicycle Archive collects documented LLM code-generated pelicans across dates, from SVG to animation, 3D and interactive works. Actual outputs, original sources, successes and failures — kept together.

Live charts ↗ · Badges read the live API and may be cached. The timeline is a subset; benchmarks stay separate.

One prompt. A whole flock.

Generate an SVG of a pelican riding a bicycle.

Images only · Timeline · Play · Sources · Benchmarks

Related MCP server: Paper Chaser MCP

The earlier answers

Original media and credits stay with each record. Model labels are source-reported, not independently authenticated. Collection method.

All works sorts by source-recorded artwork date, newest first. Timeline keeps model-month groups, prioritizes the latest counted artwork within the same month, and sorts artworks within each exact version by date. Documented releases and explicitly provisional earliest-source-work months retain their original evidence. Artwork dates are never replaced, and no pending UI section is shown. Same-model works are expanded by default with optional folding. Moving previews play silently; genuine frames add no works. Direct text-to-video is outside the main collection; preserved legacy IDs/media remain uncounted context. New imports require generationMethod=code-generated and public codeGenerationEvidence.

Explore the field guide

API & MCP

curl 'https://pelicanmap.aveniqa.com/api/v1/specimens?q=Gemini&lang=en&limit=3'
MCP  https://pelicanmap.aveniqa.com/mcp
     search_specimens · get_specimen · get_timeline

JSON · CSV · OpenAPI · Agent guide · Developer guide · Integration examples

All works/search uses artwork-date ordering by source-recorded date and stable ID. /api/v1/timeline and kind=timeline retain model-release: model month, then latestArtworkDate/latestArtworkId within the month, then artwork dates within each exact version. Latest-artwork fields express activity, not model releases. Verified releaseDate is separately sourced. Missing release evidence uses the earliest counted source-work month as an explicit inferred-position, integrated without a pending UI section; estimatedFrom keeps its evidence, not a claimed launch day. Timeline years filter effective model positions; other search years still filter artwork dates. Legacy API year=unknown selects absent verified releases. counts.timeline is a subset of counts.cases; never add the two. Complete exports retain uncounted archives and Benchmark references: inspect caseVisible, timelineVisible and referenceOnly.

Run locally

git clone https://github.com/UncleK/pelicanmap.git
cd pelicanmap
npm ci                 # Node.js 24
npm test
npm run start:api       # 127.0.0.1:48670

The checkout includes public catalog snapshots and the read-only API. Full archival builds also need the original media archive. Development · Contribute.

Collection updates preserve complete original motion: retain HTML/CSS/JS, required dependencies and original modules rather than extracting an SVG that loses its surrounding animation code. Reviewed moving works use dynamic covers in standard, compact and image-only views, with visible-only playback and static fallbacks. Restoring motion keeps the existing work ID, date, attribution and media; it does not add a work. Public intake does not require an open-source license. Source gaps remain documented; no replacement motion is invented. Reusable rules and byte-verified restorations live in collecting-policy.json and motion-restorations.json.

Original code: MIT. Artworks and upstream materials keep their original rights and licenses. Credits & licenses. Original prompt: Simon Willison.

Related MCP Connectors

Related MCP Servers

  • F
    license
    Not graded
    quality
    D
    maintenance
    Enables natural language search and exploration of over 1.15 million museum records from Museums Victoria, including specimens, items, articles, and species, with full-text search and detailed record retrieval.
    1
    -
  • A
    license
    Not graded
    quality
    D
    maintenance
    Enables AI assistants to conduct academic research workflows such as paper discovery, literature mapping, citation chasing, author pivots, citation repair, and regulatory or species document retrieval.
    MIT
  • A
    license
    A
    quality
    B
    maintenance
    Enables scientists and their AI agents to check quoted passages against pinned open-access PubMed Central full texts, look up retraction status from PMC and Crossref, and run recorded literature searches. Every quote check, search and annotation is written to an append-only, hash-chained log that can be verified, giving a dated record of what was searched and found.
    7
    MIT
  • A
    license
    A
    quality
    A
    maintenance
    Enables AI assistants to search and retrieve historical materials from five public archives in a unified format, including full-text search inside scanned books, rights statements, and provenance-tracked media downloads.
    23
    MIT