Cafecito

Project Cafecito (Espresso publication) — A subscription news publication produced under a published editorial-and-AI policy, whose byproduct is a high-frequency, machine-generated-and-reviewed stream of domain-tagged current-event records across AI, aviation, cybersecurity, careers, markets, geopolitics, robotics and startups — each item carrying an entity, a dated development and usually specific figures (contract counts and values, revenue and margin lines, yield percentages, price movements). Notably, its market coverage quantifies contract structures rather than just prices: take-or-pay …

Cafecito offers (Alternative, Sentiment, Price) — Multiple articles per day across eight tags, each roughly 9,000 characters with embedded chart imagery — this single piece runs about 9,000 characters. At visible cadence the archive accumulates on the order of a thousand or more dated, tagged, entity-attributed event records per year, each typically containing several extracted numeric claims..

High-frequency event-stream ingestion for market-monitoring and alerting systems that need breadth across domains rather than depth in one; entity-level news sentiment pipelines; structured extraction of contract-structure disclosures (prepaid volumes, floor prices, deposit backing) that most price feeds do not carry; a labelled human-in-the-loop AI-generated text corpus for model evaluation, valuable precisely because its AI provenance is disclosed at source; and policy tracking of published editorial and information-collection standards as a governance reference.

The data is with 2 years of history.

Coverage spans US, APAC; Electronic Publishing & Distribution; alternative, sentiment, price; equities.