Where queries work
The same grammar across every surface means one mental model.
Shape
Every clause is:: (not =). Combine clauses with AND, OR, and NOT:
Value types
Values are typed — autocomplete shows what’s valid per field:Autocomplete
The search input autocompletes both field names and values. As you type, it suggests:- Matching field names (with type and description)
- After
:, the valid values for that field
Verified examples
Example queries:- OS family
- Tag scope
- Combine
- Either-or
- Negation
- Vital state
os is an enum of Windows, MacOs, Linux.Vital states are booleans
A common source of confusion: Vital state isn’t a single string. Each integration exposes four boolean fields per device:
Compose them with
AND for the state you want:
Real integration prefixes
Integrations that exposesystem.* Vitals include:
system.azure_adsystem.crowdstrikesystem.jamfsystem.microsoft_defender_atpsystem.microsoft_intunesystem.onprem_adsystem.sccmsystem.servicenowsystem.zscaler_client
Saved views
Build a query → click the Save icon in the footer → name it → it appears in the view dropdown next to **Custom ***. Saving lets you reuse the query and schedule it as a report.Natural language
In AI-enabled tenants, the command bar in Device Explorer accepts plain English. The AI translates intent into a structured query you can then edit:“show me windows devices in End-User that don’t have CrowdStrike”becomes something like:
Tips
What’s not in the query language
Based on autocomplete-only verification, these patterns are not supported by the current grammar:<,>,<=,>=numeric / time comparisons (autocomplete offers value tokens, not comparators)contains,starts with,ends with(string operators)IN (...)lists — useORinsteadis empty/is not empty
: can’t express directly, build it via composition (multiple clauses with AND/OR/NOT) or use the API for richer filtering.
See also
Query Syntax (full reference)
Every confirmed field path
Tags & Logic Engine
Same syntax drives tag rules
Devices
Where you’ll spend most query time