Semantic news search - use for any news or current-events question.
params:
- query: natural language search query.
- k: articles to return (1-100).
- trust_gte: optional minimum trust (0.0-1.0).
- score_gte: minimum match score on 0-10 scale (default 4.0; set 0 to disable).
- score_lte: optional maximum match score on 0-10 scale.
- days: lookback days (default window: last 7 days).
- allow_all_dates: search the full indexed coverage window instead of days.
- sort_by: best_score (default), similarity, date_desc, date_asc.
- language: filter by language, e.g. english, hebrew, french (case-insensitive).
- country: ISO-2 country codes in UPPERCASE, e.g. US, IL, GB, DE.
- sentiment: positive, negative, or neutral (case-insensitive; omit to skip).
- category: e.g. Politics, Sport, Health (case-insensitive; maps to IPTC labels).
- domain: include only these source domains, e.g. cnn.com, yahoo.com.
- exclude_domain: exclude these source domains (must not overlap with domain).
- topic: News API topic tags (case-insensitive).
- person: named people mentioned in articles (case-insensitive).
- organization: organizations mentioned in articles (case-insensitive).
- location: places mentioned in articles (case-insensitive).
- ticker: stock ticker symbols (case-insensitive; normalized to uppercase).
- source_type: local_news, newsroom, or gov_news (case-insensitive). When the user asks for local news, local outlets, or a local news search, set source_type to ["local_news"] and do not leave that only in the query text. newsroom is for national or general newsrooms. gov_news is for government sources.
- political_bias: political bias label from trust metadata (case-insensitive).
- trust_category: trust category labels (case-insensitive).
- domain_rank_gte: minimum source domain rank (integer).
- domain_rank_lte: maximum source domain rank (integer).
- allow_multiple_chunks_per_article: when true, return multiple chunks from the same article.