APITube turns the world's news into structured JSON, so you don't have to run a crawler. It aggregates articles from 500,000+ sources across 180+ countries in 60+ languages and enriches every one before you see it: sentiment scores, extracted entities (people, organizations, locations, brands, events), IPTC categories, topics and industries.

One endpoint does the searching: GET /v1/news/everything takes 65+ filter parameters — keywords, language, country, source domain, source quality rank, sentiment range, entity IDs, media presence, date ranges — combined in a single request. The others cover the common shapes:

/v1/news/top-headlines — breaking coverage
/v1/news/story — articles clustered into one story
/v1/news/trends — trending entities, topics and categories over time buckets
/v1/news/stream — live Server-Sent Events feed
/v1/suggest/entities — turn "Tesla" into an entity ID before you filter on it

Auth is a single API key: api_key query parameter, X-API-Key header, or Authorization: Bearer. Responses are JSON by default, with JSONL, XML, CSV, TSV, XLSX and Parquet exports on paid plans. Webhooks and WebSocket dispatches push new matches to your endpoint as they get indexed. SDKs for Node.js, PHP and Python live at github.com/apitube.

Wiring up an AI agent? There is a hosted MCP server at https://mcp.apitube.io/ with search_news and suggest tools — nothing to install, just an API key in the Authorization header (paid plans).

Free tier: 100 requests/day, no card. Paid plans from $29/month. Docs: https://docs.apitube.io