Skip to main content
Well-constructed queries determine the quality of your CatchAll results. This guide shows you how to write queries that return focused and relevant results.

Quick checklist

Before submitting a query, verify it:
Start here if you’re unsure about your query. Each point is explained in detail below.

Target events in news

CatchAll searches for events in news articles, not in general knowledge databases. Your queries should describe events that appear in the news.
AI company acquisitions announced this month
FDA drug approvals for cancer treatments
Venture capital funding rounds for fintech startups
Ask yourself: “Would a journalist write a news article about this?” If not, rephrase as a specific event.

Define clear event type

Target specific, recognizable event types with clear structure:
Event CategoryExamples
BusinessAcquisitions, mergers, IPOs, funding rounds, bankruptcies
CorporateLayoffs, executive appointments, restructuring, earnings
RegulatoryFDA approvals, product recalls, legal rulings, policy changes
TechnologyProduct launches, patent filings, security breaches
MarketStock listings, delisting, major price movements
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)

Balance specificity

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.
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.

Focus on recent events

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.
Tech IPOs announced this week
Recent cybersecurity incidents at healthcare companies
AI model releases in the past month

Time-based queries for monitors

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
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:
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)
See Configure monitors for monitor-specific guidance.

Use single entity

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.
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)

See also