Common Corpus
Pleias
2.27 trillion tokens of text that is either uncopyrighted or openly licensed, with per-document provenance.
A multilingual collection of books, newspapers, scientific articles, government and legal documents, and code. Each document carries information about its source and license. It was built so models can be trained without relying on content of uncertain legal status.
Access
- Cost
- FreeNo charge.
- License
- Open licenseMixed per-document; the card says only uncopyrighted or freely licensed data is included
- Sign-up
- No sign-up
- How you get it
- API and Bulk downloadParquet
What's in it
- One row is
- One row per document, with provenance metadata
- Covers
- Mostly Europe and North America; English and French dominate
- Years
- Spans centuries (digitized books) to recent web and code
- Updated
- Irregular
- Size
- 2.27 trillion tokens
Quality
Documented provenance and an ICLR 2026 paper. Old OCR text and uneven source coverage limit its use as a general corpus.
- OCR errors in historical texts
- Per-document license must be checked for specific uses
- English and French dominate other languages
Connects to
Questions it could answer
- How much of the corpus is public-domain versus explicitly licensed?
- Which collections supply the non-English text?