TEXT
COMMERCIAL
RAW
Pick something you have used, write at least 80 words on what you thought and why; copied text, reviews under 80 words and reviews of unused things are rejected. Three per contributor at most, from a global base strongest in Southeast Asia, the United States and Europe, so products, services and expectations come from several markets, not one marketplace. No star ratings or aspect labels; the reasoning supports aspect extraction well. Scraped review sets carry academic-only or unclear terms and cannot say whether a human wrote the review; these were typed in-app by people who opted in to this use. A consented human baseline for review-authenticity detectors as much as a sentiment corpus.
Volume
10,000+
Reviews
Contributors
5,000+
Regions
100+ countries
Updated
What’s in the collection and how it’s delivered.
Data modality
Text
Annotated
Raw (no labels)
Exclusivity
Non-exclusive rights
Capture tool
Typed in the Kled app
Use cases
Sentiment and opinion summarization, preference modeling with clear provenance, review-authenticity baselines
Format
Plain text
Resolution
Plain text, UTF-8
Total volume
10,000+ reviews (approx. 50 MB)
Regions
100+ countries
Purity level
99% passed QA review
Where this data comes from and how it’s licensed.
Every file comes straight from the contributor who provided it and is rights-cleared before it’s listed, so your license traces back to the source.
Contributor
Provides the data and uploads it straight to Kled.
Kled
Clears the rights before anything is listed.
Your team
Licenses it under the terms below.
License terms
Non-exclusive commercial training license.
Consent
Each contributor opted in to this use when they uploaded.
Specs reflect the current collection and may change as new data is added. Final terms are set in your license agreement.
