{"service":"attention-index","positioning":"An attention/curiosity index over Wikipedia readership and the global news stream. Explicitly NOT Google Trends and NOT Google search interest — Wikipedia pageviews are a curiosity proxy, and GDELT is editorial news coverage.","sources":[{"name":"Wikimedia Pageviews API","api":"wikimedia.org/api/rest_v1/metrics/pageviews (+ MediaWiki Action search)","homepage":"https://www.wikimedia.org/","license":"CC0 / public-domain pageview aggregates + article titles (article TEXT is CC BY-SA and is NEVER returned by this service)"},{"name":"GDELT Project","api":"GDELT DOC 2.0 TimelineVol (reused read-only)","homepage":"https://www.gdeltproject.org/","license":"unlimited/unrestricted use incl. commercial, redistribution permitted"}],"license":"Wikimedia pageview aggregates: CC0 / public-domain factual data, free reuse incl. commercial (article TEXT stays CC BY-SA and is never returned). GDELT: unlimited/unrestricted use incl. commercial, redistribution permitted.","attribution":"Attention index computed from Wikipedia pageview aggregates (Wikimedia REST Pageviews API + MediaWiki search, CC0 factual counts + article titles; article text is NOT rehosted) and the global news stream via the GDELT Project (public-domain). This is an attention/curiosity index over Wikipedia readership and world news coverage — NOT Google Trends and NOT Google search interest.","projects":["ar.wikipedia","de.wikipedia","en.wikipedia","es.wikipedia","fr.wikipedia","hi.wikipedia","it.wikipedia","ja.wikipedia","ko.wikipedia","pt.wikipedia","ru.wikipedia","zh.wikipedia"],"project_aliases":["ar","arabic","chinese","de","en","english","es","fr","french","german","hi","hindi","it","italian","ja","japanese","ko","korean","portuguese","pt","ru","russian","spanish","zh"],"defaults":{"project":"en.wikipedia","window_days":90,"min_window_days":14,"max_window_days":365,"spike_k":3.0},"params":{"keyword":"free-text word/entity (ticker, company, product, person, topic)","project":"Wikipedia language (en, es, fr, de, ru, ja, ...); default en.wikipedia","window / days":"lookback in days (14-365, default 90)","article":"explicit Wikipedia article title to skip word->article resolution"},"rate_limit_note":"Wikimedia asks for a contactful User-Agent (a burst without one gets 429). This service sends one and enforces a global serialized throttle + a ~6h TTL cache, so repeated keyword polling is served from cache.","notes":"The word->article resolver is REQUIRED: a raw ticker like NVDA search-matches a 4-views/day decoy, so the top candidates are disambiguated by real readership and the rejected ones are returned under resolved.alternatives (honesty about what was measured). Brand keywords often redirect to the substance article (Ozempic -> Semaglutide). Multi-word / non-entity keywords with no article fall back to the GDELT news-volume axis (wikipedia_available: false), never a 500."}