Skip to main content
Search filters are useful ways to find exactly what you need across all company documents. Instead of browsing hundreds of results to find the single document you are looking for, use filters to say “show me only Google Drive presentations from Jane updated last week.” You can use filter to:
  • Perform detailed and specific queries in Glean Search.
  • Create Manual search queries for Agent, allowing fine-grained specification of exactly which data Company Search, Web Search, and other Agents Actions/Steps will retrieve.
Basic syntax: filter:value (like app:gdrive or from:"Jane Doe"). Include negation filters by adding a - operator to any query (e.g. -app:gmail excludes results from Google Mail). Combine filters to get precise results — it’s like giving Glean specific instructions to understand exactly what you’re looking for. This comprehensive article is designed to be your go-to reference for Glean Search Filters:

​
Essential syntax rules

The essential syntax rules include how to handle spaces, exact phrases, and word matching logic. Quote multi-word values, use exact phrases for precise matches, and understand that unquoted words search for ANY while quoted words search for ALL. 
filter:value                    → Basic filter syntax
filter:"value with space"       → ALWAYS quote multi-word values  
"exact phrase"                  → Find this exact phrase
"word1" "word2"                 → Find ALL these words (any order)
word1 word2                     → Find ANY of these words
-filter:value                   → EXCLUDE certain results.

​
Core filters

These are the fundamental filters that work across all applications: 
app:appname                    → Limit to specific app(s)
type:typename                  → Filter by document type  
from:person                    → From specific person
from:me                        → Your documents
updated:timeframe              → Recently modified
my:history                     → You've viewed (6 month limit)
collection:"Name"              → In a Glean collection
has:golink                     → Has a Go link
in:"Folder Name"               → Docs in specific container/space/folder

​
Date and time expressions

​
Natural language

updated:today                  → Current day only
updated:yesterday              → Previous day only
updated:past_week              → Last 7 days
updated:past_month             → Last 30 days
updated:past_3_days            → Last 3 days
updated:past_quarter           → Last 90 days
updated:this_week              → Current calendar week
updated:last_month             → Previous calendar month
after:"2024-01-01"             → After a specific date
before:"2024-01-01"             → Before a specific date

​
Specific dates

updated:"2024-01-15"                    → Specific date
updated:"2024-01"                       → Entire month
updated:"2024"                          → Entire year
updated:"2024-Q1"                       → Quarter
after:"2024-01-01" before:"2024-01-31"  → Date range

​
Boolean logic rules

  • Same filter = OR: type:doc type:sheet (docs OR sheets).
  • Different filters = AND: app:jira from:me (Jira AND from me).
  • No way to force AND on same filter (platform limitation).
  • No parentheses or custom grouping (platform limitation).
  • Negation filter: Add a - operator to remove certain keywords from queries (e.g. employee experience -customer → exclude results that mention “customer”.).

​
People and entity filters

Filters for finding people by title, department, location, reporting structure, and combining with document searches. 
# Job & Organization
title:"Software Engineer"      → Job title
department:engineering         → Department
businessunit:sales             → Business unit
team:"Platform Team"           → Specific team

# Location
location:"San Francisco"       → City/office
country:"United States"        → Country
state:"CA"                     → State/province
region:"EMEA"                  → Region

# Reporting & Employment
reportsto:"Manager Name"       → Direct manager
level:[ic|manager|director]    → Seniority level
startafter:"2023-01-01"        → Start date after
startbefore:"2024-01-01"       → Start date before
employmenttype:[fulltime|contractor|intern]

# Combining with documents
from:[title:"Product Manager"] type:presentation
from:[department:sales] updated:past_week

​
Text matching patterns

# Exact phrase
"quarterly business review"     → Exact phrase in order

# All words required
"quarterly" "business" "review" → All 3 words, any order

# Any words
quarterly business review       → ANY of these words

# Special characters (use quotes)
"user@company.com"
"Q4-2024"  
"$1,000,000"
"C++"

​
Use search filters in Glean Agents

Crafting manual search queries in Glean Agents is a powerful approach for achieving accurate, highly relevant information retrieval across organizational knowledge. Unlike AI-generated queries, manually constructed searches allow you to specify exactly which parameters and filters to apply, leading to precise results that align tightly with business objectives. Quoting multi-word values—such as account:“Acme Corp” ensures the system interprets search intent correctly, while explicit filter inclusion helps target the right data set from the start. This level of control not only minimizes noise and irrelevant results but also supports reproducibility and transparency in automated processes. Manual queries allows you to confidently surface critical insights, troubleshoot with nuance, and create efficient workflows that are robust to context changes—making them indispensable for high-stakes use cases within Glean. Hence, for the Company Search step in a Glean Agent, it is a core best practice to default to a Manual search query instead of an AI-generated one. Manual queries provide deterministic, repeatable, and reliable results, which is critical for any automated workflow.
Use the AI-generated option only when search parameters, like a specific date, must be created dynamically during the Agent’s run.

​
Critical Rules

Always quote multi-word values, test syntax in Glean Search UI first, and be explicit with all needed filters for reliable Agent results.
  • Always quote multi-word values: account:"Acme Corp".
  • Test in Glean Search UI first, copy exact syntax.
    Any manual search query that will work in Agents will also work in Glean Search. Consequently, this pattern of testing manual queries is quite effective.
  • Use Manual Search (not AI-generated) when possible.
  • Be explicit - include all needed filters.
  • Handle empty results in your Agent logic.

​
Good vs Bad Examples

Proper quoting makes the difference between successful and failed searches. Quoted multi-word values ensure accurate interpretation while unquoted ones cause parsing errors. 
✅ GOOD: account:"Acme Corporation" status:"in progress"
❌ BAD:  account:Acme Corporation status:in progress

✅ GOOD: app:jira type:bug label:"high-priority" 
❌ BAD:  app:jira type:bug label:high priority

​
Tips and tricks

Build queries incrementally starting simple, use app filters first for performance, and discover new filters through the UI’s “All filters” option.

​
Query Building Strategy

The incremental layering approach - starting with keywords, then systematically adding app, type, person, and time filters - mirrors how Glean’s search engine processes queries for optimal performance. This methodology prevents the “too many results” problem that overwhelms users and ensures each filter layer meaningfully refines the result set rather than creating contradictory constraints.
  1. Start simple: project phoenix
  2. Add app: project phoenix app:confluence
  3. Add type: project phoenix app:confluence type:page
  4. Add person: project phoenix app:confluence type:page from:"PM"
  5. Add time: project phoenix app:confluence type:page from:"PM" updated:past_week

​
Performance Tips

Begin with app filters to narrow scope, be specific with document types, limit date ranges, and use exact phrases for known titles.
  • Use app: filter first to narrow scope
  • Be specific with types
  • Limit date ranges when possible
  • Use exact phrases for known titles

​
Discover filters through the Search UI:

  1. Run a search.
  2. Click All filters.
  3. Select options.
  4. Copy the syntax from the search bar.

​
Known limitations and workarounds

LimitationWhat doesn’t workPractical workaround
No complex boolean(A OR B) AND (C OR D)Run multiple targeted searches
No same‑field AND“Docs with Jane AND John”Use text search for both names
No regexPattern matchingUse multiple exact phrases
History scopemy:history > 6 monthsCombine with other filters
GDrive recency edge casesNewly updated files may not appear with broad time filtersCombine time with from: or type:; add keywords for critical searches

​
Quick reference table

These are the 8 common use cases with their exact filter combinations for quick copy-paste:
NeedFilter combination
My recent workfrom:me updated:past_week
Team documentsfrom:[department:"Dept"] type:document
Open bugsapp:jira type:bug status:"open"
Customer intel"Acme Corp" app:salescloud app:zendesk
Meeting notes"meeting" type:document updated:past_month
Shared drivesapp:gdrive folder:"Shared"
Recent Slackapp:slack updated:yesterday
My history, slidesmy:history type:presentation

​
Common search patterns

These are ready-to-use queries for team documents, customer intelligence, project tracking, bugs, sales, and knowledge search. Recent team documents 
from:[department:"Engineering"] updated:past_week
from:[reportsto:"Manager Name"] "meeting notes"
from:me app:gdrive type:document updated:past_month
Customer intelligence 
"Acme Corp" app:salescloud app:zendesk app:gong
account:"Acme Corp" type:opportunity status:"open"
"Acme Corp" app:slack channel:"#customer-success"
Project tracking 
"Project Phoenix" app:confluence app:jira
"Project Phoenix" updated:past_week -type:email
"Project Phoenix" from:[team:"Phoenix Team"]
Bug management 
app:jira type:bug status:"open" priority:highest
app:jira type:bug label:customer-reported assignee:me
app:jira type:bug status:"resolved" updated:past_week
Sales pipeline 
app:salescloud type:opportunity stage:!"closed lost"
app:salescloud type:opportunity amount:>100000
app:salescloud closedate:"2024-Q1" owner:me
Knowledge search 
"how to" type:document app:confluence label:tutorial
collection:"Engineering Playbooks" "best practices"
has:golink "onboarding"

​
Application specific filters

The following are detailed filter syntax for Google Drive, Slack, Jira, Confluence, Salesforce, GitHub/GitLab, Teams, Gong, and Zendesk. Google Drive 
app:gdrive
type:[document|spreadsheet|presentation|pdf|image|video|folder]
folder:"Folder Name"
owner:email@company.com
# Example:
app:gdrive type:spreadsheet folder:"Finance" updated:past_week
Slack 
app:slack
channel:"#channel-name"   # include the #
from:@username            # include the @
type:[conversation|channel|dm]
# Example:
app:slack channel:"#engineering" from:@john "deployment"
Jira 
app:jira
type:[bug|story|epic|task|subtask]
status:"Status Name"
assignee:"Person Name"
reporter:"Person Name"
project:"PROJECT-KEY"
label:labelname
component:componentname
priority:[highest|high|medium|low|lowest]
sprint:"Sprint Name"
# Example:
app:jira type:bug status:"in progress" priority:high assignee:me
Confluence 
app:confluence
type:[page|blogpost|space]
space:"Space Name"
author:"Person Name"
label:labelname
# Example:
app:confluence space:"Engineering Wiki" author:"Jane" label:howto
Salesforce 
# Sales Cloud
app:salescloud
type:[opportunity|account|lead|contact|case]
status:"Status Name"
stage:"Stage Name"
owner:"Person Name"
account:"Account Name"
amount:>100000
closedate:"2024-Q1"

# Service Cloud
app:servicecloud
type:case
priority:[high|medium|low]
# Example:
app:salescloud type:opportunity stage:"negotiation" amount:>50000
GitHub / GitLab 
app:github   # or app:gitlab
type:[issue|pr|commit|repo|discussion]
repo:"repository-name"
author:"username"
assignee:"username"
label:labelname
status:[open|closed|merged]
# Example:
app:github type:pr repo:"main-app" status:open label:bug
Microsoft Teams 
app:teams
team:"Team Name"
channel:"Channel Name"
from:"Person Name"
type:[conversation|file|channel]
# Example:
app:teams team:"Product Dev" channel:"General" type:file
Gong 
app:gong
account:"Account Name"
participant:"Person Name"
type:[call|email]
# Example:
app:gong account:"Acme Corp" type:call updated:past_week
Zendesk 
app:zendesk
type:ticket
status:[new|open|pending|solved|closed]
priority:[urgent|high|normal|low]
assignee:"Agent Name"
requester:"Customer Name"
# Example:
app:zendesk type:ticket status:open priority:urgent