Assemble the corpus
Import materials by keyword, URL, CSV export, or raw text
Enter any keywords. Article URLs are collected from news search, then scraped at the edge (up to 25 per round).
Optional date range. Leave both fields empty to scrape without a date filter. Rows without a usable date are always retained and counted.
One URL per line. The first 25 are fetched per round.
Corpora produced by SemanText: Datetime, Title, Text, URL, TextID, Publication.
Paste raw text or upload plain-text files (.txt). Each submission is treated as a single document and is limited to 100,000 tokens.
Corpus overview
All imported materials, in one place
02No corpus yet. Assemble one in step 01: import articles by keyword or URL, or add raw text.
Analysis tools
All computation belongs to your browser, not to a server