Cyoc: 100M-1B Scale Interactive Stories from Cyoc.net

Cyoc contains 100 million to 1 billion units of English-language narrative text scraped from the interactive fiction site cyoc.net. Created by tgdatasets and updated in March 2026, the collection preserves user-contributed story chapters archived via the Wayback Machine. It focuses on the 'Choose Your Own Change' community's branching story paths.

Use Cases

• Training language models on branching narrative structures in story chapters
• Fine-tuning text generation for interactive fiction styles
• Analyzing community-driven narrative tropes from the cyoc.net archive

Strengths

• 100M to 1B size category for narrative text
• Preserves niche interactive fiction from the cyoc.net community
• English language focus

Limitations

• Broken narrative paths due to incomplete scraping
• Potential for messy text extraction from HTML
• Unknown license status

Provenance

Source

cyoc.net via Wayback Machine

Collection Method

scraped

Freshness

Last updated March 2026

Geography

United States