About Outerwoven
Based in Bend, Oregon
Outerwoven tracks the words and phrases spiking in public language. It notices when a word suddenly rises, in what people publish and in what they look up, then explains what it means now and shows where it started and how it spread. Today it reads news outlets, podcasts, newsletters and YouTube, plus public lookup data, and it adds sources over time. The stories and conversations around a phrase are the evidence behind it.
Modern linguistics
Outerwoven practises modern linguistics: the study of language as it is actually used, measured from real, dated sources rather than from what a dictionary says it should mean. Words change meaning, move between communities and pass from niche to mainstream. Outerwoven records that as it happens and keeps the record, so how a word changed can be traced later.
Words often rise because of serious or painful events. Outerwoven describes how a word is being used; it does not rank events or make light of what a word refers to.
What the Outerwoven software does
- Collects text and media. Four times a day Outerwoven gathers new articles, episodes, newsletter posts and videos from the sources it follows. Each item links back to its original publisher; Outerwoven does not claim ownership of that content and shows only the original headline, a short snippet and a link, never the full text.
- Counts words and phrases. Software counts the single words and two- and three-word phrases that appear each day in each kind of source, and records where and when each phrase was first seen. For podcasts that publish their own transcripts, the transcript is counted too; transcripts are used only for counting and are never shown. Every phrase page charts that use over time, and notes when English Wiktionary first added an entry for the phrase.
- Explains what phrases mean. A large language model writes a short plain-English meaning for a phrase, based only on how the phrase is used in the items Outerwoven has collected. Meanings are written by software, not by an editor, and they never name people. Beside it, phrase pages show the phrase's English Wiktionary definition, credited and shared under CC BY-SA, so the dictionary sense and the current use can be compared.
- Finds rising words. Outerwoven compares each phrase with its own recent average to show which are growing fastest. To focus on language rather than the names in the news, it leaves out names of people, places and organizations and everyday English words, using word frequencies from the Google Books Ngram dataset (CC BY 3.0) and the list of entries in English Wiktionary (CC BY-SA). The archive is young, so this is labelled Early.
- Groups coverage into conversations. Machine learning language models compare the meaning of each item's headline, so coverage of the same story from a newspaper, a podcast and a video lands in one conversation, even when the headlines share few words. A language model gives each conversation a neutral name that describes the topic, never a person, and sorts it into a category such as Politics & World, Business & Finance, Science & Technology, Entertainment & Culture, Sports, Health, or Lifestyle & Society. Original publisher headlines are shown as published, attributed and linked.
- Reads your text. On Read a text, anyone can paste an article or a paragraph and run the same phrase-finding software on it. It marks the words Outerwoven follows and lists the conversations using them, with a meaning where one has been written. The text is not saved.
- Estimates reach. For each conversation it combines figures the platforms themselves publish, such as views, subscribers and engagement, into a clearly labelled estimate of how many people it reached.
Runs on its own
No editor picks or orders the words or stories. Software gathers, counts, groups, labels and ranks everything on a schedule.
Follow Outerwoven
Using Outerwoven
Outerwoven is free. Anyone can browse the phrase pages and the top conversations, no account needed; a free account lets you follow phrases. See the Terms of Use and Privacy Policy for details.
Contact
Outerwoven is based in Bend, Oregon. Questions or feedback? Email contact@outerwoven.com.