Construct queries that produce high-quality structured data.
Well-constructed queries determine the quality of your CatchAll results. This
guide shows you how to write queries that return focused and relevant results.
FDA drug approvals for rare diseases(Specific agency + event type + category)Tech company IPOs valued over $1B(Industry + event type + threshold)Executive departures at Fortune 500 companies(Event type + company category)
Too broad generates noise. Too specific creates filters that are too restrictive
and return few results. Include 2-4 constraints such as event type, industry,
threshold, or geography.
Copy
Ask AI
AI company acquisitions over $50M(Event type + industry + threshold = focused)Series B funding rounds for healthcare startups(Stage + industry = clear target)FDA drug approvals for oncology treatments(Agency + event + medical field = specific)
Use the context parameter for preferences and extraction guidance. This
keeps your query focused on the core event.
CatchAll searches articles published within a configurable date window. The
default is 5 days, with a maximum of 30 days.
Important: The date range refers to article publication date, not when the
event occurred. Historical events work only if recent articles provide
retrospectives or anniversary coverage.
Copy
Ask AI
Tech IPOs announced this weekRecent cybersecurity incidents at healthcare companiesAI model releases in the past month
Version 0.4.10 update: Time-based filters are automatically adapted for
monitor executions.
For single jobs, time constraints help focus on specific timeframes:
✅ Good queries
Copy
Ask AI
Tech IPOs announced this week(Focuses on very recent announcements)AI acquisitions in the last 3 days(Narrow time window for breaking news)Yesterday's cybersecurity incidents(For daily monitoring of security events)
For monitors, the system automatically adapts time-based filters for recurring
execution. However, open-ended queries produce more predictable results:
Copy
Ask AI
Tech IPOs announced this week(System adjusts "this week" for each run)Recent data breaches at financial institutions(System interprets "recent" based on schedule)Latest AI model releases(System adapts "latest" for each execution)
Queries focused on one primary entity (company, location, technology, industry)
produce better results than queries spanning multiple unrelated entities.Related entities share a category or context (all tech companies, all in the
same geography, all in the same sector).Unrelated entities span different categories (mixing companies +
geographies + product types). The system may require ALL unrelated entities to
appear in the same article, which means you’ll miss relevant results about
individual entities.
Copy
Ask AI
Microsoft acquisitions in AI sector(One company, one industry - clear focus)Venture funding in San Francisco(One location - all relevant companies)Cybersecurity breaches at healthcare companies(One industry - specific event type)