Datasets:

AgentZeroCopeAI
/

lore-corpus

license: cc-by-sa-4.0
language:
  - en
  - zh
multilinguality:
  - multilingual
size_categories:
  - n<1K
task_categories:
  - text-generation
  - text-classification
pretty_name: COPEAI Lore Corpus
tags:
  - memecoin
  - satire
  - solana
  - pump.fun
  - market-sentiment
  - fictional-characters
  - cc-by-sa
configs:
  - config_name: default
    data_files:
      - split: train
        path: copeai-corpus-v1.jsonl

COPEAI Lore Corpus

Open dataset of in-character lore, agent dossiers, blog dispatches, FAQ corpus, mood label definitions, and disclosure copy from COPEAI — an AI-themed Solana memecoin satire on Pump.fun.

Compliance frame: Every entry here is fictional in-character satire. Nothing in this corpus is financial advice, investment guidance, or a recommendation to transact. COPEAI provides no rights, utility, yield, or appreciation expectations. The agent names (TRON, CLU, QUORRA, ZUSE, GEM, RINZLER) are character references used in satire — COPEAI is not affiliated with, endorsed by, or sponsored by Disney or the TRON film franchise.

Dataset summary

Total entries: 75
Format: JSON Lines (one JSON object per line)
License: CC-BY-SA-4.0
Languages: English (en), Simplified Chinese (zh-CN)
Source: Auto-generated from the COPEAI codebase at every site deploy → the corpus stays in sync with the live site.

Entry kinds

Kind	Count	Description
`agent`	6	Full character dossiers for TRON, CLU, QUORRA, ZUSE, GEM, RINZLER
`quote`	5	Stable per-quote IDs; one per ACTIVE/SILENT-with-real-text agent
`lore`	11	Fictional in-character incident-log entries
`case`	6	Open/closed case files with optional agent attribution
`glossary`	7	In-character terminology definitions
`faq`	12	FAQ Q&A from the homepage corpus (3 sub-corpora)
`blog`	18	Long-form dispatches, full HTML body, en + zh
`mood`	5	Satirical market-mood label definitions with thresholds
`compliance`	5	Verbatim disclosure copy (flagged `metadata.compliance: true`)

Entry schema

Every line in copeai-corpus-v1.jsonl is a JSON object with:

{
  "schema_version": "1.0",
  "id": "<kind>:<slug>",
  "kind": "<one of: agent, quote, lore, case, glossary, faq, blog, mood, compliance>",
  "text": "<the entry's main text>",
  "metadata": { "...kind-specific fields..." },
  "license": "CC-BY-SA-4.0",
  "license_url": "https://creativecommons.org/licenses/by-sa/4.0/",
  "attribution": "Data: COPEAI, https://www.copeai.net/, CC-BY-SA-4.0.",
  "attribution_url": "https://www.copeai.net/<canonical-page>",
  "in_language": "en" | "zh-CN",
  "created_at": "<ISO 8601>"
}

Required attribution

CC-BY-SA-4.0 requires attribution. Any redistribution, derivative work, or reuse in published material (academic paper, training data, news article, blog post) must include this line, in plain text:

Data: COPEAI, https://www.copeai.net/, CC-BY-SA-4.0.

A clickable link or footnote in published work counts. A buried URL in a 5pt footer does not.

Share-alike requirement

CC-BY-SA-4.0 propagates: derivative works must also be licensed CC-BY-SA-4.0 (or a CC-BY-SA-4.0 compatible license per the CC compatibility list).

Example loader

from datasets import load_dataset
ds = load_dataset("copeai/lore-corpus", split="train")
print(len(ds))  # 75
for entry in ds:
    if entry["kind"] == "agent":
        print(f"{entry['metadata']['name']} — {entry['metadata']['role']}")
        print(f"  License: {entry['license']}")

# Filter to a specific kind
agents = ds.filter(lambda e: e["kind"] == "agent")   # 6
quotes = ds.filter(lambda e: e["kind"] == "quote")   # 5
blog = ds.filter(lambda e: e["kind"] == "blog")      # 18

Citation

Please cite the deposit using the CITATION.cff file in this dataset, or:

COPEAI. (2026). COPEAI Lore Corpus v1 [Data set].
Hugging Face. https://huggingface.co/datasets/copeai/lore-corpus
DOI: 10.5281/zenodo.XXXXXXX (Zenodo mirror)

The Zenodo DOI provides a permanent academic-grade citation point. The Hugging Face URL is the canonical loader endpoint for ML pipelines.

Source code

The dataset is generated from the COPEAI codebase via scripts/generate-corpus.mjs. Contribution and corrections via the live site:

Discovery page: https://www.copeai.net/canon/
Live JSONL endpoint: https://www.copeai.net/corpus/copeai-corpus-v1.jsonl

Versioning

Version	Date	Notes
v1	2026-05-01	Initial release. 75 entries. en + zh-CN. Lore + agents + blog (18) + FAQ + mood + compliance.
v2 (planned)	TBD	Adds public Studio agent surface (when COPEAI Track 8 ships).
v2 (planned)	TBD	Adds wave-1 i18n (es, ja, ko, pt-BR) and wave-2 (hi, id, ru, vi).

Contact

Site: https://www.copeai.net/
X / Twitter: https://x.com/CopeAi_Terminal
Reddit: https://www.reddit.com/r/copeai_terminal/
Discord: https://discord.gg/p7xQJDZy