Datasets for RAG, fine-tuning and training built from public sources
AI writing that doesn't sound like AI
Stage
Not given
Not given
Founded
Not given
Not given
Based in
Not given
Not given
Pricing model
Paid
Freemium
Pricing
Web Corpus: starter sets from $49 (250,000-500,000 documents); shelf corpora from $99-$199 per shelf (one-time license); Everything (full crawl) $3,000 one-time; Whole Library (all builds for one year) $5,000 per year. Reference Collections: Encyclopedia $1,200 commercial or $349 research (one-time)
Free plan with 2 requests per day; Basic $15/mo (200 generations/month); Pro $49/mo (800 generations/month); Premium $99/mo (2,000 generations/month); API: $2.50 per million input tokens, $12 per million output and thinking tokens, rounded up to nearest cent with 1 cent minimum per request.
Key features
Web corpus with 100.7M documents and 105B tokens from Common Crawl
Documents scored with Bear Rank and shelved by subject
15 million encyclopedia entries with sources and model attribution
200 million synthetic Q&A records with metadata
Delivered as JSONL and Parquet formats
Free samples available on Hugging Face
Command-line client with resumable downloads and SHA-256 verification
Statistics collection with 7,000 topics and 140,000 figures
Glossaries for 10,000 topics with one million definitions
Book Q&A dataset with 8,000 titles for reading comprehension
Distribution fine-tuning for natural writing output
Browser-based console for generating and rewriting
Server-to-server API with /v1/generate endpoint
Specialized agents for SEO, rewriting, and copyediting
Custom model training on company style
Generation history and public share links
AI advising for AI transformation and strategy
Lightweight metrics and checking tools in editor
What makes it different
All datasets built only from public sources with full provenance tracking
Every record includes source attribution and licensing information
Eight curated subject shelves for targeted corpus selection
Flexible purchasing: buy individual shelves, slices, or entire library
Over 4 TB of clean, deduplicated text ready for machine learning
Trained on human prose to eliminate repetitive AI patterns and 'LLM slop'
More creative and varied output than general-purpose AI assistants
Custom models trained on your company's specific writing style
Preserves author's intent and voice without hallucinations
Usage-based API pricing with no charges for failed generations
Deft