Indexing API
- Documents
- Permissions
- Troubleshooting
- Datasources
- Authentication
- People
- Shortcuts
Client API
- Activity
- Announcements
- Answers
- Authentication
- Calendar
- Chat
- POSTDetect and answer questions
- POSTChat
- POSTDeletes all saved Chats owned by a user
- POSTDeletes saved Chats
- POSTRetrieves a Chat
- POSTRetrieves all saved Chats
- POSTGets the metadata for a custom Chat application
- POSTUpload files for Chat.
- POSTGet files uploaded by a user for Chat.
- POSTDelete files uploaded by a user for chat.
- POST
- Agents
- Collections
- Displayable Lists
- Documents
- Images
- Insights
- Messages
- Pins
- Search
- Entities
- Shortcuts
- Summarize
- Tools
- User
- Verification
Actions API
- Authentication
- Setup
Chat
Chat
Chat
Chat
Have a conversation with Glean AI.
curl --request POST \
--url https://{domain}-be.glean.com/rest/api/v1/chat \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"messages": [
{
"author": "USER",
"messageType": "CONTENT",
"fragments": [
{
"text": "What are the company holidays this year?"
}
]
}
]
}'{
"messages": [
{
"author": "GLEAN_AI",
"messageType": "CONTENT",
"hasMoreFragments": false,
"agentConfig": {
"agent": "DEFAULT",
"mode": "DEFAULT"
},
"fragments": [
{
"text": "There are no holidays!"
}
]
}
]
}Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Headers
Email address of a user on whose behalf the request is intended to be made (should be non-empty only for global tokens).
Auth type being used to access the endpoint (should be non-empty only for global tokens).
Query Parameters
The offset of the client's timezone in minutes from UTC. e.g. PDT is -420 because it's 7 hours behind UTC.
Body
application/json
Includes chat history for Glean AI to respond to.
A list of chat messages, from most recent to least recent. It can be assumed that the first chat message in the list is the user's most recent query.
A message that is rendered as one coherent unit with one given sender.
Describes the agent config that generated this message. Populated on responses and not required on requests.
Available options:
USER, GLEAN_AI A list of Citations that were used to generate the response.
Information about the source for a ChatMessage.
An opaque token that represents this particular result in this particular ChatMessage. To be used for /feedback reporting.
The Glean Document ID.
The app or other repository type from which the document was extracted
The source from which document content was pulled, e.g. an API crawl or browser history
Available options:
API_CRAWL, BROWSER_CRAWL, BROWSER_HISTORY, BUILTIN, FEDERATED_SEARCH, PUSH_API, WEB_CRAWL, NATIVE_HISTORY The datasource-specific type of the document (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
The plaintext content of the document.
The title of the document.
A permalink for the document.
The datasource instance from which the document was extracted.
The type of the result. Interpretation is specific to each datasource. (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
The name of the container (higher level parent, not direct parent) of the result. Interpretation is specific to each datasource (e.g. Channels for Slack, Project for Jira). cf. parentId
The Glean Document ID of the container. Uniquely identifies the container.
The Glean Document ID of the super container. Super container represents a broader abstraction that contains many containers. For example, whereas container might refer to a folder, super container would refer to a drive.
The id of the direct parent of the result. Interpretation is specific to each datasource (e.g. parent issue for Jira). cf. container
The index-wide unique identifier.
A unique identifier used to represent the document in any logging or feedback requests in place of documentId.
Hash of the Glean Document ID.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
A list of people mentioned in the document.
The level of visibility of the document as understood by our system.
Available options:
PRIVATE, SPECIFIC_PEOPLE_AND_GROUPS, DOMAIN_LINK, DOMAIN_VISIBLE, PUBLIC_LINK, PUBLIC_VISIBLE A list of components this result is associated with. Interpretation is specific to each datasource. (e.g. for Jira issues, these are components.)
The status or disposition of the result. Interpretation is specific to each datasource. (e.g. for Jira issues, this is the issue status such as Done, In Progress or Will Not Fix).
The status category of the result. Meant to be more general than status. Interpretation is specific to each datasource.
A list of stars associated with this result. "Pin" is an older name.
The document priority. Interpretation is datasource specific.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
A list of tags for the document. Interpretation is datasource specific.
A list of collections that the document belongs to.
The user-visible datasource specific id (e.g. Salesforce case number for example, GitHub PR number).
A list of shortcuts of which destination URL is for the document.
For file datasources like onedrive/github etc this has the path to the file
Custom fields specific to individual datasources
The document's document_category(.proto).
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
A thumbnail image representing this document.
A list of documents that are ancestors of this document in the hierarchy of the document's datasource, for example parent folders or containers. Ancestors can be of different types and some may not be indexed. Higher level ancestors appear earlier in the list.
Example:
{
"container": "container",
"parentId": "JIRA_EN-1337",
"createTime": "2000-01-23T04:56:07.000Z",
"datasource": "datasource",
"author": { "name": "name" },
"documentId": "documentId",
"updateTime": "2000-01-23T04:56:07.000Z",
"mimeType": "mimeType",
"objectType": "Feature Request",
"components": ["Backend", "Networking"],
"status": ["Done"],
"customData": { "someCustomField": "someCustomValue" }
}
A list of content sub-sections in the document, e.g. text blocks with different headings in a Drive doc or Confluence page.
Structure for file uploaded by a user for Chat.
Unique identifier of the file.
Example:
"FILE_1234"
Url of the file.
Example:
"www.google.com"
Name of the uploaded file.
Example:
"sample.pdf"
Metadata of a file uploaded by a user for Chat.
Current status of the file.
Available options:
PROCESSING, PROCESSED, FAILED, DELETED Upload time, in epoch seconds.
Size of the processed file in bytes.
Reason for failed status.
Available options:
PARSE_FAILED, AV_SCAN_FAILED, FILE_TOO_SMALL, FILE_TOO_LARGE, FILE_EXTENSION_UNSUPPORTED, FILE_METADATA_VALIDATION_FAIL, FILE_PROCESSING_TIMED_OUT MIME type of the file.
The display name.
An opaque identifier that can be used to request metadata for a Person.
A list of documents related to this person.
How this document relates to the including entity.
Available options:
ATTACHMENT, CANONICAL, CASE, CONTACT, CONVERSATION_MESSAGES, EXPERT, FROM, HIGHLIGHT, OPPORTUNITY, RECENT, SOURCE, TICKET, TRANSCRIPT, WITH Which entity in the response that this entity relates to. Relevant when there are multiple entities associated with the response (such as merged customers)
Example:
{
"query": "app:github type:pull author:mortimer",
"label": "Mortimer's PRs",
"datasource": "github"
}
A truncated list of documents with this relation. TO BE DEPRECATED.
A truncated list of documents associated with this relation. To be used in favor of documents because it contains a trackingToken.
Available options:
FULL_TIME, CONTRACTOR, NON_EMPLOYEE, FORMER_EMPLOYEE Example:
"FULL_TIME"
The first name of the person
The last name of the person
Job title.
Typically the highest level organizational unit; generally applies to bigger companies with multiple distinct businesses.
An organizational unit where everyone has a similar task, e.g. Engineering.
Info about the employee's team(s).
Use id if you index teams via Glean, and use name and externalLink if you want to use your own team pages
The number of people in this person's department.
The user's primary email address
Additional email addresses of this user beyond the primary, if any.
User facing string representing the person's location.
Detailed location with information about country, state, city etc.
Link to a customer's internal profile page. This is set to '#' when no link is desired.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The chain of reporting in the company as far up as it goes. The last entry is this person's direct manager.
Phone number as a number string.
The timezone of the person. E.g. "Pacific Daylight Time".
The offset of the person's timezone in seconds from UTC.
The URL of the person's avatar. Public, glean-authenticated and Base64 encoded data URLs are all valid (but not third-party-authenticated URLs).
The original photo URL of the person's avatar before any edits they made are applied
The URL of the person's banner photo.
The date when the employee started.
If a former employee, the last date of employment.
Short biography or mission statement of the employee.
She/her, He/his or other pronoun.
The total recursive size of the people reporting to this person, or 1
The total number of people who directly report to this person, or 0
The preferred name of the person, or a nickname.
List of social network profiles.
List of profiles this user has in different datasources / tools that they use.
List of people and distances to those people from this person. Optionally with metadata.
Information regarding the invite status of a person.
Whether the user has signed into Glean at least once.
The last time the user has used the Glean extension in ISO 8601 format.
Describes the permissions levels that a user has for permissioned features. When the client sends this, Permissions.read and Permissions.write are the additional permissions granted to a user on top of what they have via their roles. When the server sends this, Permissions.read and Permissions.write are the complete (merged) set of permissions the user has, and Permissions.roles is just for display purposes.
User customizable fields for additional people information.
The logging id of the person used in scrubbed logs, tracking GA metrics.
Percentage of the company that started strictly after this person. Between [0,100).
Intervals of busy time for this person, along with the type of event they're busy with.
A generic, light-weight calendar event.
flag settings to indicate user profile settings for certain items
The badges that a user has earned over their lifetime.
Displays a user's accomplishment or milestone
Whether this person is a "root" node in their organization's hierarchy.
Example:
{
"department": "Movies",
"email": "george@example.com",
"location": "Hollywood, CA",
"phone": 6505551234,
"photoUrl": "https://example.com/george.jpg",
"startDate": "2000-01-23",
"title": "Actor"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Each reference range and its corresponding snippets
Each text range from the response can correspond to an array of snippets from the citation source.
A subsection of a given string to which some special formatting should be applied.
The inclusive start index of the range.
The exclusive end index of the range.
Available options:
BOLD, CITATION, LINK The URL associated with the range, if applicable. For example, the linked URL for a LINK range.
A document corresponding to the range, if applicable. For example, the cited document for a CITATION range.
A matching snippet from the document. Query term matches are marked by the unicode characters uE006 and uE007.
The mime type of the snippets, currently either text/plain or text/html.
A matching snippet from the document with no highlights.
Used for sorting based off the snippet's location within all_snippetable_text
The bolded ranges within text.
A subsection of a given string to which some special formatting should be applied.
A URL, generated based on availability, that links to the position of the snippet text or to the nearest header above the snippet text.
IDs of files uploaded in the message that are referenced to generate the answer.
A list of rich data used to represent the response or formulate a request. These are linearly stitched together to support richer data formats beyond simple text.
Represents a part of a ChatMessage that originates from a single action/tool. It is designed to support rich data formats beyond simple text, allowing for a more dynamic and interactive chat experience. Each fragment can include various types of content, such as text, search queries, action information, and more. Also, each ChatMessageFragment should only have one of structuredResults, querySuggestion, writeAction, followupAction, or file.
The search queries issued while responding.
The query being suggested (e.g. enforcing the missing term from the original query).
A query term missing from the original query on which this suggestion is based.
Information about the search provider that generated this suggestion.
Name of the search provider.
URL to the provider's logo.
URL template that can be used to perform the suggested search by replacing the {query} placeholder with the query suggestion.
Example:
{
"name": "Google",
"logo": "https://app.glean.com/images/feather/globe.svg",
"searchLinkUrlTemplate": "https://www.google.com/search?q={query}&hl=en"
}
A user-facing description to display for the suggestion.
The datasource associated with the suggestion.
The maximum number of FacetBuckets to return in each FacetResult.
Filter results to a single datasource name (e.g. gmail, slack). All results are returned if missing.
Filter results to one or more datasources (e.g. gmail, slack). All results are returned if missing.
If true, the operators in the query are taken to override any operators in facetFilters in the case of conflict. This is used to correctly set rewrittenFacetFilters and rewrittenQuery.
A list of filters for the query. An AND is assumed between different facetFilters. For example, owner Sumeet and type Spreadsheet shows documents that are by Sumeet AND are Spreadsheets.
A list of facet filter sets that will be OR'ed together. SearchRequestOptions where both facetFilterSets and facetFilters set are considered as bad request. Callers should set only one of these fields.
Within a single FacetFilterSet, the filters are treated as AND. For example, owner Sumeet and type Spreadsheet shows documents that are by Sumeet AND are Spreadsheets.
Facets for which FacetResults should be fetched and that don't apply to a particular datasource. If specified, these values will replace the standard default facets (last_updated_at, from, etc.). The requested facets will be returned alongside datasource-specific facets if searching a single datasource.
Auth tokens which may be used for non-indexed, federated results (e.g. Gmail).
Hints that the QE should return result counts (via the datasource facet result) for all supported datasources, rather than just those specified in the datasource[s]Filter
Array of hints containing which fields should be populated in the response.
Hints for the response content.
Available options:
ALL_RESULT_COUNTS, FACET_RESULTS, QUERY_METADATA, RESULTS, SPELLCHECK_METADATA The offset of the client's timezone in minutes from UTC. e.g. PDT is -420 because it's 7 hours behind UTC.
Whether or not to disable spellcheck.
Disables automatic adjustment of the input query for spelling corrections or other reasons.
[beta] Enables expanded content to be returned for LLM usage. The size of content per result returned should be modified using maxSnippetSize. Server may return less or more than what is specified in maxSnippetSize. For more details, https://docs.google.com/document/d/1CTOLSxWWT9WDEnHVLoCUaxbGYyXYP8kctPRF-RluSQY/edit. Requires sufficient permissions.
A list of filters which restrict the search results to only the specified content.
A list of filters specifying content to avoid getting search results from. Exclusions take precendence over inclusions and other query parameters, such as search operators and search facets.
Example:
{
"datasourceFilter": "JIRA",
"datasourcesFilter": ["JIRA"],
"queryOverridesFacetFilters": true,
"facetFilters": [
{
"fieldName": "fieldName",
"values": ["fieldValues", "fieldValues"]
},
{
"fieldName": "fieldName",
"values": ["fieldValues", "fieldValues"]
}
]
}
The bolded ranges within the query of the QuerySuggestion.
A subsection of a given string to which some special formatting should be applied.
The inclusive start index of the range.
The exclusive end index of the range.
Available options:
BOLD, CITATION, LINK The URL associated with the range, if applicable. For example, the linked URL for a LINK range.
A document corresponding to the range, if applicable. For example, the cited document for a CITATION range.
Whether the associated query was at least partially copy-pasted. If subsequent requests are issued after a copy-pasted query is constructed (e.g. with facet modifications), this bit should continue to be set for those requests.
Example:
{ "hasCopyPaste": true }
Example:
{
"query": "app:github type:pull author:mortimer",
"label": "Mortimer's PRs",
"datasource": "github"
}
Files referenced in the message fragment. This is used to construct rich-text messages with file references.
Unique identifier of the file.
Example:
"FILE_1234"
Url of the file.
Example:
"www.google.com"
Name of the uploaded file.
Example:
"sample.pdf"
Metadata of a file uploaded by a user for Chat.
Current status of the file.
Available options:
PROCESSING, PROCESSED, FAILED, DELETED Upload time, in epoch seconds.
Size of the processed file in bytes.
Reason for failed status.
Available options:
PARSE_FAILED, AV_SCAN_FAILED, FILE_TOO_SMALL, FILE_TOO_LARGE, FILE_EXTENSION_UNSUPPORTED, FILE_METADATA_VALIDATION_FAIL, FILE_PROCESSING_TIMED_OUT MIME type of the file.
Basic information about an action. This can be used to construct rich-text messages with action references.
The manifest for a tool that can be used to augment Glean Assistant.
The type of tool.
Available options:
RETRIEVAL, ACTION Unique identifier for the tool. Name should be understandable by the LLM, and will be used to invoke a tool.
Human understandable name of the tool. Max 50 characters.
Description of the tool meant for a human.
An opaque id which is unique identifier for the tool.
URL used to fetch the logo.
Name of the generated object. This will be used to indicate to the end user what the generated object contains.
Example:
["HR ticket", "Email", "Chat message"]
Indicates the kind of knowledge a tool would access or modify.
Available options:
NEUTRAL_KNOWLEDGE, COMPANY_KNOWLEDGE, WORLD_KNOWLEDGE The time the tool was created in ISO format (ISO 8601)
The time the tool was last updated in ISO format (ISO 8601)
Valid only for write actions. Represents the type of write action. REDIRECT - The client renders the URL which contains information for carrying out the action. EXECUTION - Send a request to an external server and execute the action.
Available options:
REDIRECT, EXECUTION The type of authentication being used. Use 'OAUTH_*' when Glean calls an external API (e.g., Jira) on behalf of a user to obtain an OAuth token. 'OAUTH_ADMIN' utilizes an admin token for external API calls on behalf all users. 'OAUTH_USER' uses individual user tokens for external API calls. 'DWD' refers to domain wide delegation.
Available options:
NONE, OAUTH_USER, OAUTH_ADMIN, API_KEY, BASIC_AUTH, DWD Config for tool's authentication method.
Usage instructions for the LLM to use this action.
Whether this action has been fully configured and validated.
An array of entities in the work graph retrieved via a data request.
A single object that can support any object in the work graph. Only a single object will be populated.
The Glean Document ID.
The app or other repository type from which the document was extracted
The source from which document content was pulled, e.g. an API crawl or browser history
Available options:
API_CRAWL, BROWSER_CRAWL, BROWSER_HISTORY, BUILTIN, FEDERATED_SEARCH, PUSH_API, WEB_CRAWL, NATIVE_HISTORY The datasource-specific type of the document (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
The title of the document.
A permalink for the document.
Example:
{
"container": "container",
"parentId": "JIRA_EN-1337",
"createTime": "2000-01-23T04:56:07.000Z",
"datasource": "datasource",
"author": { "name": "name" },
"documentId": "documentId",
"updateTime": "2000-01-23T04:56:07.000Z",
"mimeType": "mimeType",
"objectType": "Feature Request",
"components": ["Backend", "Networking"],
"status": ["Done"],
"customData": { "someCustomField": "someCustomValue" }
}
A list of content sub-sections in the document, e.g. text blocks with different headings in a Drive doc or Confluence page.
The display name.
An opaque identifier that can be used to request metadata for a Person.
A list of documents related to this person.
Example:
{
"department": "Movies",
"email": "george@example.com",
"location": "Hollywood, CA",
"phone": 6505551234,
"photoUrl": "https://example.com/george.jpg",
"startDate": "2000-01-23",
"title": "Actor"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Unique identifier.
Link to company's associated website domains.
A map of {string, int} pairs representing counts of each document type associated with this customer.
A list of POC for company.
A list of Customers.
The date when the interaction with customer started.
Average contract annual revenue with that customer.
User facing (potentially generated) notes about company.
Example:
"CIO is interested in trying out the product."
Unique identifier
Team name
A description of the team
Typically the highest level organizational unit; generally applies to bigger companies with multiple distinct businesses.
An organizational unit where everyone has a similar task, e.g. Engineering.
A link to the team's photo.
A link to the team's banner photo.
Link to a team page on the internet or your company's intranet
The members on this team
Metadata about the relationship of a person to a team.
Number of members on this team (recursive; includes all individuals that belong to this team, and all individuals that belong to a subteam within this team)
The emails for this team
The datasource profiles of the team
the data source of the team, e.g. GDRIVE
For teams created from docs, the doc title. Otherwise empty.
when this team was last updated.
whether this team is fully processed or there are still unprocessed operations that'll affect it
Available options:
PROCESSED, QUEUED_FOR_CREATION, QUEUED_FOR_DELETION can this team be deleted. Some manually ingested teams like GCS_CSV or PUSH_API cannot
The logging id of the team used in scrubbed logs, client analytics, and metrics.
A list of objects related to a source object.
The permissions the current viewer has with respect to a particular object.
Unique identifier.
Title or name of the custom entity.
The datasource the custom entity is from.
The type of the entity. Interpretation is specific to each datasource
A list of user roles for the custom entity explicitly granted by the owner.
The permissions the current viewer has with respect to a particular object.
The opaque ID of the Answer.
Example:
3
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The time the answer was created in ISO format (ISO 8601).
The time the answer was last updated in ISO format (ISO 8601).
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The parent board this answer is in.
The collections to which the answer belongs.
The document's document_category(.proto).
Glean Document ID of the Answer. The Glean Document ID is supported for cases where the Answer ID isn't available. If both are available, using the Answer ID is preferred.
Example:
"ANSWERS_answer_3"
Example:
"Why is the sky blue?"
Additional ways of phrasing this question.
The plain text answer to the question.
Example:
"From https://en.wikipedia.org/wiki/Diffuse_sky_radiation, the sky is blue because blue light is more strongly scattered than longer-wavelength light."
The parent board ID of this Answer, or 0 if it's a floating Answer.
Filters which restrict who should see the answer. Values are taken from the corresponding filters in people search.
A list of user roles for the answer added by the owner.
A list of user roles for the answer removed by the owner.
A list of roles for this answer explicitly granted by an owner, editor, or admin.
Available options:
DOCUMENT, ASSISTANT The permissions the current viewer has with respect to a particular object.
Heading text that was matched to produce this result.
Question text that was matched to produce this result.
Example:
{
"snippets": [
{
"snippet": "snippet",
"mimeType": "mimeType"
}
],
"metadata": {
"container": "container",
"createTime": "2000-01-23T04:56:07.000Z",
"datasource": "datasource",
"author": { "name": "name" },
"documentId": "documentId",
"updateTime": "2000-01-23T04:56:07.000Z",
"mimeType": "mimeType",
"objectType": "objectType"
},
"title": "title",
"url": "https://example.com/foo/bar",
"nativeAppUrl": "slack://foo/bar",
"mustIncludeSuggestions": [
{
"missingTerm": "container",
"query": "container"
}
]
}
The attendee list, including their response status
The app or other repository type this represents
The datasource-specific type of the document (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
Mimetype is used to differentiate between sub applications from a datasource (e.g. Sheets, Docs from Gdrive)
If there is available icon URL.
The unique ID of the Collection.
The unique name of the Collection.
A brief summary of the Collection's contents.
The emoji icon of this Collection.
Indicates whether edits are allowed for everyone or only admins.
The parent of this Collection, or 0 if it's a top-level Collection.
The datasource type this Collection can hold.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The number of items currently in the Collection. Separated from the actual items so we can grab the count without items.
The number of children Collections. Separated from the actual children so we can grab the count without children.
The items in this Collection.
Metadata having what categories this Collection is pinned to and the eligible categories to pin to
The names of the shortcuts (Go Links) that point to this Collection.
The children Collections of this Collection.
A list of user roles for the Collection.
A list of added user roles for the Collection.
A list of removed user roles for the Collection.
Filters which restrict who should see this Collection. Values are taken from the corresponding filters in people search.
The permissions the current viewer has with respect to a particular object.
The unique ID of the Answer Board.
The unique name of the Collection.
A brief summary of the Collection's contents.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The number of items currently in the Answer Board. Separated from the actual items so we can grab the count without items.
A list of user roles for the Answer Board.
A list of added user roles for the Collection.
A list of removed user roles for the Collection.
Filters which restrict who should see this Collection. Values are taken from the corresponding filters in people search.
The permissions the current viewer has with respect to a particular object.
Last file match for a repo
Example:
{
"repoName": "scio",
"fileName": "README.md",
"matches": [
{
"lineNumber": 1,
"content": "Welcome to the beginning",
"ranges": []
},
{
"lineNumber": 2,
"content": "Second line of the file",
"ranges": []
},
{
"lineNumber": 3,
"content": "hello world hello world",
"ranges": [
{ "startindex": 0, "endIndex": 5 },
{ "startIndex": 12, "endIndex": 17 }
]
}
]
}
Link text following go/ prefix as entered by the user.
canonical link text following go/ prefix where hyphen/underscore is removed.
Title for the Go Link
A list of user roles for the Go Link.
The opaque id of the user generated content.
Destination URL for the shortcut.
Glean Document ID for the URL, if known.
A short, plain text blurb to help people understand the intent of the shortcut.
Whether this shortcut is unlisted or not. Unlisted shortcuts are visible to author + admins only.
For variable shortcuts, contains the URL template; note, destinationUrl contains default URL.
A list of user roles added for the Shortcut.
A list of user roles removed for the Shortcut.
The permissions the current viewer has with respect to a particular object.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The time the shortcut was created in ISO format (ISO 8601).
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The time the shortcut was updated in ISO format (ISO 8601).
Document that corresponds to the destination URL, if applicable.
The URL from which the user is then redirected to the destination URL. Full replacement for https://go/<inputAlias>.
The part of the shortcut preceding the input alias when used for showing shortcuts to users. Should end with "/". e.g. "go/" for native shortcuts.
Indicates whether a shortcut is native or external.
The URL using which the user can access the edit page of the shortcut.
A list of documents related to this structured result.
How this document relates to the including entity.
Available options:
ATTACHMENT, CANONICAL, CASE, CONTACT, CONVERSATION_MESSAGES, EXPERT, FROM, HIGHLIGHT, OPPORTUNITY, RECENT, SOURCE, TICKET, TRANSCRIPT, WITH Which entity in the response that this entity relates to. Relevant when there are multiple entities associated with the response (such as merged customers)
Example:
{
"query": "app:github type:pull author:mortimer",
"label": "Mortimer's PRs",
"datasource": "github"
}
A truncated list of documents with this relation. TO BE DEPRECATED.
A truncated list of documents associated with this relation. To be used in favor of documents because it contains a trackingToken.
The text of the related question
The answer for the related question
Subsections of the answer string to which some special formatting should be applied (eg. bold)
A subsection of a given string to which some special formatting should be applied.
A disambiguation between multiple entities with the same name
Name of the ambiguous entity
The unique id of the entity in the knowledge graph
The type of entity.
Available options:
PERSON, PROJECT, CUSTOMER Any snippets associated to the populated object.
A matching snippet from the document. Query term matches are marked by the unicode characters uE006 and uE007.
The mime type of the snippets, currently either text/plain or text/html.
A matching snippet from the document with no highlights.
Used for sorting based off the snippet's location within all_snippetable_text
The bolded ranges within text.
A subsection of a given string to which some special formatting should be applied.
A URL, generated based on availability, that links to the position of the snippet text or to the nearest header above the snippet text.
An opaque token that represents this particular result in this particular query. To be used for /feedback reporting.
The level of visual distinction that should be given to a result.
Available options:
HERO, PROMOTED, STANDARD Source context for this result. Possible values depend on the result type.
Available options:
EXPERT_DETECTION, ENTITY_NLQ An opaque token that represents this particular result in this particular query. To be used for /feedback reporting.
Response timestamp of the message.
A unique server-side generated ID used to identify a message, automatically populated for any USER authored messages.
Opaque tracking token generated server-side.
Semantically groups content of a certain type. It can be used for purposes such as differential UI treatment. USER authored messages should be of type CONTENT and do not need messageType specified.
Available options:
UPDATE, CONTENT, CONTEXT, DEBUG, DEBUG_EXTERNAL, ERROR, HEADING, WARNING Signals there are additional response fragments incoming.
Save the current interaction as a Chat for the user to access and potentially continue later.
The id of the Chat that context should be retrieved from and messages added to. An empty id starts a new Chat, and the Chat is saved if saveChat is true.
A list of filters which only allows chat to access certain content.
Specifications for containers that should be used as part of the restriction (include/exclude). Memberships are recursively defined for a subset of datasources (currently: SharePoint, OneDrive, Google Drive, and Confluence). Please contact the Glean team to enable this for more datasources. Recursive memberships do not apply for Collections.
The URL of the document.
A list of filters which disallows chat from accessing certain content. If content is in both inclusions and exclusions, it'll be excluded.
Specifications for containers that should be used as part of the restriction (include/exclude). Memberships are recursively defined for a subset of datasources (currently: SharePoint, OneDrive, Google Drive, and Confluence). Please contact the Glean team to enable this for more datasources. Recursive memberships do not apply for Collections.
The URL of the document.
Timeout in milliseconds for the request. A 408 error will be returned if handling the request takes longer.
Example:
30000
The ID of the application this request originates from, used to determine the configuration of underlying chat processes. This should correspond to the ID set during admin setup. If not specified, the default chat experience will be used.
If set, response lines will be streamed one-by-one as they become available. Each will be a ChatResponse, formatted as JSON, and separated by a new line. If false, the entire response will be returned at once. Note that if this is set and the model being used does not support streaming, the model's response will not be streamed, but other messages from the endpoint still will be.
Response
200
text/plain
OK
A single response from the /chat backend.
A message that is rendered as one coherent unit with one given sender.
Describes the agent config that generated this message. Populated on responses and not required on requests.
Available options:
USER, GLEAN_AI A list of Citations that were used to generate the response.
Information about the source for a ChatMessage.
An opaque token that represents this particular result in this particular ChatMessage. To be used for /feedback reporting.
The Glean Document ID.
The app or other repository type from which the document was extracted
The source from which document content was pulled, e.g. an API crawl or browser history
Available options:
API_CRAWL, BROWSER_CRAWL, BROWSER_HISTORY, BUILTIN, FEDERATED_SEARCH, PUSH_API, WEB_CRAWL, NATIVE_HISTORY The datasource-specific type of the document (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
The plaintext content of the document.
The title of the document.
A permalink for the document.
The datasource instance from which the document was extracted.
The type of the result. Interpretation is specific to each datasource. (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
The name of the container (higher level parent, not direct parent) of the result. Interpretation is specific to each datasource (e.g. Channels for Slack, Project for Jira). cf. parentId
The Glean Document ID of the container. Uniquely identifies the container.
The Glean Document ID of the super container. Super container represents a broader abstraction that contains many containers. For example, whereas container might refer to a folder, super container would refer to a drive.
The id of the direct parent of the result. Interpretation is specific to each datasource (e.g. parent issue for Jira). cf. container
The index-wide unique identifier.
A unique identifier used to represent the document in any logging or feedback requests in place of documentId.
Hash of the Glean Document ID.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
A list of people mentioned in the document.
The level of visibility of the document as understood by our system.
Available options:
PRIVATE, SPECIFIC_PEOPLE_AND_GROUPS, DOMAIN_LINK, DOMAIN_VISIBLE, PUBLIC_LINK, PUBLIC_VISIBLE A list of components this result is associated with. Interpretation is specific to each datasource. (e.g. for Jira issues, these are components.)
The status or disposition of the result. Interpretation is specific to each datasource. (e.g. for Jira issues, this is the issue status such as Done, In Progress or Will Not Fix).
The status category of the result. Meant to be more general than status. Interpretation is specific to each datasource.
A list of stars associated with this result. "Pin" is an older name.
The document priority. Interpretation is datasource specific.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
A list of tags for the document. Interpretation is datasource specific.
A list of collections that the document belongs to.
The user-visible datasource specific id (e.g. Salesforce case number for example, GitHub PR number).
A list of shortcuts of which destination URL is for the document.
For file datasources like onedrive/github etc this has the path to the file
Custom fields specific to individual datasources
The document's document_category(.proto).
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
A thumbnail image representing this document.
A list of documents that are ancestors of this document in the hierarchy of the document's datasource, for example parent folders or containers. Ancestors can be of different types and some may not be indexed. Higher level ancestors appear earlier in the list.
Example:
{
"container": "container",
"parentId": "JIRA_EN-1337",
"createTime": "2000-01-23T04:56:07.000Z",
"datasource": "datasource",
"author": { "name": "name" },
"documentId": "documentId",
"updateTime": "2000-01-23T04:56:07.000Z",
"mimeType": "mimeType",
"objectType": "Feature Request",
"components": ["Backend", "Networking"],
"status": ["Done"],
"customData": { "someCustomField": "someCustomValue" }
}
A list of content sub-sections in the document, e.g. text blocks with different headings in a Drive doc or Confluence page.
Structure for file uploaded by a user for Chat.
Unique identifier of the file.
Example:
"FILE_1234"
Url of the file.
Example:
"www.google.com"
Name of the uploaded file.
Example:
"sample.pdf"
Metadata of a file uploaded by a user for Chat.
Current status of the file.
Available options:
PROCESSING, PROCESSED, FAILED, DELETED Upload time, in epoch seconds.
Size of the processed file in bytes.
Reason for failed status.
Available options:
PARSE_FAILED, AV_SCAN_FAILED, FILE_TOO_SMALL, FILE_TOO_LARGE, FILE_EXTENSION_UNSUPPORTED, FILE_METADATA_VALIDATION_FAIL, FILE_PROCESSING_TIMED_OUT MIME type of the file.
The display name.
An opaque identifier that can be used to request metadata for a Person.
A list of documents related to this person.
How this document relates to the including entity.
Available options:
ATTACHMENT, CANONICAL, CASE, CONTACT, CONVERSATION_MESSAGES, EXPERT, FROM, HIGHLIGHT, OPPORTUNITY, RECENT, SOURCE, TICKET, TRANSCRIPT, WITH Which entity in the response that this entity relates to. Relevant when there are multiple entities associated with the response (such as merged customers)
Example:
{
"query": "app:github type:pull author:mortimer",
"label": "Mortimer's PRs",
"datasource": "github"
}
A truncated list of documents with this relation. TO BE DEPRECATED.
A truncated list of documents associated with this relation. To be used in favor of documents because it contains a trackingToken.
Available options:
FULL_TIME, CONTRACTOR, NON_EMPLOYEE, FORMER_EMPLOYEE Example:
"FULL_TIME"
The first name of the person
The last name of the person
Job title.
Typically the highest level organizational unit; generally applies to bigger companies with multiple distinct businesses.
An organizational unit where everyone has a similar task, e.g. Engineering.
Info about the employee's team(s).
Use id if you index teams via Glean, and use name and externalLink if you want to use your own team pages
The number of people in this person's department.
The user's primary email address
Additional email addresses of this user beyond the primary, if any.
User facing string representing the person's location.
Detailed location with information about country, state, city etc.
Link to a customer's internal profile page. This is set to '#' when no link is desired.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The chain of reporting in the company as far up as it goes. The last entry is this person's direct manager.
Phone number as a number string.
The timezone of the person. E.g. "Pacific Daylight Time".
The offset of the person's timezone in seconds from UTC.
The URL of the person's avatar. Public, glean-authenticated and Base64 encoded data URLs are all valid (but not third-party-authenticated URLs).
The original photo URL of the person's avatar before any edits they made are applied
The URL of the person's banner photo.
The date when the employee started.
If a former employee, the last date of employment.
Short biography or mission statement of the employee.
She/her, He/his or other pronoun.
The total recursive size of the people reporting to this person, or 1
The total number of people who directly report to this person, or 0
The preferred name of the person, or a nickname.
List of social network profiles.
List of profiles this user has in different datasources / tools that they use.
List of people and distances to those people from this person. Optionally with metadata.
Information regarding the invite status of a person.
Whether the user has signed into Glean at least once.
The last time the user has used the Glean extension in ISO 8601 format.
Describes the permissions levels that a user has for permissioned features. When the client sends this, Permissions.read and Permissions.write are the additional permissions granted to a user on top of what they have via their roles. When the server sends this, Permissions.read and Permissions.write are the complete (merged) set of permissions the user has, and Permissions.roles is just for display purposes.
User customizable fields for additional people information.
The logging id of the person used in scrubbed logs, tracking GA metrics.
Percentage of the company that started strictly after this person. Between [0,100).
Intervals of busy time for this person, along with the type of event they're busy with.
A generic, light-weight calendar event.
flag settings to indicate user profile settings for certain items
The badges that a user has earned over their lifetime.
Displays a user's accomplishment or milestone
Whether this person is a "root" node in their organization's hierarchy.
Example:
{
"department": "Movies",
"email": "george@example.com",
"location": "Hollywood, CA",
"phone": 6505551234,
"photoUrl": "https://example.com/george.jpg",
"startDate": "2000-01-23",
"title": "Actor"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Each reference range and its corresponding snippets
Each text range from the response can correspond to an array of snippets from the citation source.
A subsection of a given string to which some special formatting should be applied.
The inclusive start index of the range.
The exclusive end index of the range.
Available options:
BOLD, CITATION, LINK The URL associated with the range, if applicable. For example, the linked URL for a LINK range.
A document corresponding to the range, if applicable. For example, the cited document for a CITATION range.
A matching snippet from the document. Query term matches are marked by the unicode characters uE006 and uE007.
The mime type of the snippets, currently either text/plain or text/html.
A matching snippet from the document with no highlights.
Used for sorting based off the snippet's location within all_snippetable_text
The bolded ranges within text.
A subsection of a given string to which some special formatting should be applied.
A URL, generated based on availability, that links to the position of the snippet text or to the nearest header above the snippet text.
IDs of files uploaded in the message that are referenced to generate the answer.
A list of rich data used to represent the response or formulate a request. These are linearly stitched together to support richer data formats beyond simple text.
Represents a part of a ChatMessage that originates from a single action/tool. It is designed to support rich data formats beyond simple text, allowing for a more dynamic and interactive chat experience. Each fragment can include various types of content, such as text, search queries, action information, and more. Also, each ChatMessageFragment should only have one of structuredResults, querySuggestion, writeAction, followupAction, or file.
The search queries issued while responding.
The query being suggested (e.g. enforcing the missing term from the original query).
A query term missing from the original query on which this suggestion is based.
Information about the search provider that generated this suggestion.
Name of the search provider.
URL to the provider's logo.
URL template that can be used to perform the suggested search by replacing the {query} placeholder with the query suggestion.
Example:
{
"name": "Google",
"logo": "https://app.glean.com/images/feather/globe.svg",
"searchLinkUrlTemplate": "https://www.google.com/search?q={query}&hl=en"
}
A user-facing description to display for the suggestion.
The datasource associated with the suggestion.
The maximum number of FacetBuckets to return in each FacetResult.
Filter results to a single datasource name (e.g. gmail, slack). All results are returned if missing.
Filter results to one or more datasources (e.g. gmail, slack). All results are returned if missing.
If true, the operators in the query are taken to override any operators in facetFilters in the case of conflict. This is used to correctly set rewrittenFacetFilters and rewrittenQuery.
A list of filters for the query. An AND is assumed between different facetFilters. For example, owner Sumeet and type Spreadsheet shows documents that are by Sumeet AND are Spreadsheets.
A list of facet filter sets that will be OR'ed together. SearchRequestOptions where both facetFilterSets and facetFilters set are considered as bad request. Callers should set only one of these fields.
Within a single FacetFilterSet, the filters are treated as AND. For example, owner Sumeet and type Spreadsheet shows documents that are by Sumeet AND are Spreadsheets.
Facets for which FacetResults should be fetched and that don't apply to a particular datasource. If specified, these values will replace the standard default facets (last_updated_at, from, etc.). The requested facets will be returned alongside datasource-specific facets if searching a single datasource.
Auth tokens which may be used for non-indexed, federated results (e.g. Gmail).
Hints that the QE should return result counts (via the datasource facet result) for all supported datasources, rather than just those specified in the datasource[s]Filter
Array of hints containing which fields should be populated in the response.
Hints for the response content.
Available options:
ALL_RESULT_COUNTS, FACET_RESULTS, QUERY_METADATA, RESULTS, SPELLCHECK_METADATA The offset of the client's timezone in minutes from UTC. e.g. PDT is -420 because it's 7 hours behind UTC.
Whether or not to disable spellcheck.
Disables automatic adjustment of the input query for spelling corrections or other reasons.
[beta] Enables expanded content to be returned for LLM usage. The size of content per result returned should be modified using maxSnippetSize. Server may return less or more than what is specified in maxSnippetSize. For more details, https://docs.google.com/document/d/1CTOLSxWWT9WDEnHVLoCUaxbGYyXYP8kctPRF-RluSQY/edit. Requires sufficient permissions.
A list of filters which restrict the search results to only the specified content.
A list of filters specifying content to avoid getting search results from. Exclusions take precendence over inclusions and other query parameters, such as search operators and search facets.
Example:
{
"datasourceFilter": "JIRA",
"datasourcesFilter": ["JIRA"],
"queryOverridesFacetFilters": true,
"facetFilters": [
{
"fieldName": "fieldName",
"values": ["fieldValues", "fieldValues"]
},
{
"fieldName": "fieldName",
"values": ["fieldValues", "fieldValues"]
}
]
}
The bolded ranges within the query of the QuerySuggestion.
A subsection of a given string to which some special formatting should be applied.
The inclusive start index of the range.
The exclusive end index of the range.
Available options:
BOLD, CITATION, LINK The URL associated with the range, if applicable. For example, the linked URL for a LINK range.
A document corresponding to the range, if applicable. For example, the cited document for a CITATION range.
Whether the associated query was at least partially copy-pasted. If subsequent requests are issued after a copy-pasted query is constructed (e.g. with facet modifications), this bit should continue to be set for those requests.
Example:
{ "hasCopyPaste": true }
Example:
{
"query": "app:github type:pull author:mortimer",
"label": "Mortimer's PRs",
"datasource": "github"
}
Files referenced in the message fragment. This is used to construct rich-text messages with file references.
Unique identifier of the file.
Example:
"FILE_1234"
Url of the file.
Example:
"www.google.com"
Name of the uploaded file.
Example:
"sample.pdf"
Metadata of a file uploaded by a user for Chat.
Current status of the file.
Available options:
PROCESSING, PROCESSED, FAILED, DELETED Upload time, in epoch seconds.
Size of the processed file in bytes.
Reason for failed status.
Available options:
PARSE_FAILED, AV_SCAN_FAILED, FILE_TOO_SMALL, FILE_TOO_LARGE, FILE_EXTENSION_UNSUPPORTED, FILE_METADATA_VALIDATION_FAIL, FILE_PROCESSING_TIMED_OUT MIME type of the file.
Basic information about an action. This can be used to construct rich-text messages with action references.
The manifest for a tool that can be used to augment Glean Assistant.
The type of tool.
Available options:
RETRIEVAL, ACTION Unique identifier for the tool. Name should be understandable by the LLM, and will be used to invoke a tool.
Human understandable name of the tool. Max 50 characters.
Description of the tool meant for a human.
An opaque id which is unique identifier for the tool.
URL used to fetch the logo.
Name of the generated object. This will be used to indicate to the end user what the generated object contains.
Example:
["HR ticket", "Email", "Chat message"]
Indicates the kind of knowledge a tool would access or modify.
Available options:
NEUTRAL_KNOWLEDGE, COMPANY_KNOWLEDGE, WORLD_KNOWLEDGE The time the tool was created in ISO format (ISO 8601)
The time the tool was last updated in ISO format (ISO 8601)
Valid only for write actions. Represents the type of write action. REDIRECT - The client renders the URL which contains information for carrying out the action. EXECUTION - Send a request to an external server and execute the action.
Available options:
REDIRECT, EXECUTION The type of authentication being used. Use 'OAUTH_*' when Glean calls an external API (e.g., Jira) on behalf of a user to obtain an OAuth token. 'OAUTH_ADMIN' utilizes an admin token for external API calls on behalf all users. 'OAUTH_USER' uses individual user tokens for external API calls. 'DWD' refers to domain wide delegation.
Available options:
NONE, OAUTH_USER, OAUTH_ADMIN, API_KEY, BASIC_AUTH, DWD Config for tool's authentication method.
Usage instructions for the LLM to use this action.
Whether this action has been fully configured and validated.
An array of entities in the work graph retrieved via a data request.
A single object that can support any object in the work graph. Only a single object will be populated.
The Glean Document ID.
The app or other repository type from which the document was extracted
The source from which document content was pulled, e.g. an API crawl or browser history
Available options:
API_CRAWL, BROWSER_CRAWL, BROWSER_HISTORY, BUILTIN, FEDERATED_SEARCH, PUSH_API, WEB_CRAWL, NATIVE_HISTORY The datasource-specific type of the document (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
The title of the document.
A permalink for the document.
Example:
{
"container": "container",
"parentId": "JIRA_EN-1337",
"createTime": "2000-01-23T04:56:07.000Z",
"datasource": "datasource",
"author": { "name": "name" },
"documentId": "documentId",
"updateTime": "2000-01-23T04:56:07.000Z",
"mimeType": "mimeType",
"objectType": "Feature Request",
"components": ["Backend", "Networking"],
"status": ["Done"],
"customData": { "someCustomField": "someCustomValue" }
}
A list of content sub-sections in the document, e.g. text blocks with different headings in a Drive doc or Confluence page.
The display name.
An opaque identifier that can be used to request metadata for a Person.
A list of documents related to this person.
Example:
{
"department": "Movies",
"email": "george@example.com",
"location": "Hollywood, CA",
"phone": 6505551234,
"photoUrl": "https://example.com/george.jpg",
"startDate": "2000-01-23",
"title": "Actor"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Unique identifier.
Link to company's associated website domains.
A map of {string, int} pairs representing counts of each document type associated with this customer.
A list of POC for company.
A list of Customers.
The date when the interaction with customer started.
Average contract annual revenue with that customer.
User facing (potentially generated) notes about company.
Example:
"CIO is interested in trying out the product."
Unique identifier
Team name
A description of the team
Typically the highest level organizational unit; generally applies to bigger companies with multiple distinct businesses.
An organizational unit where everyone has a similar task, e.g. Engineering.
A link to the team's photo.
A link to the team's banner photo.
Link to a team page on the internet or your company's intranet
The members on this team
Metadata about the relationship of a person to a team.
Number of members on this team (recursive; includes all individuals that belong to this team, and all individuals that belong to a subteam within this team)
The emails for this team
The datasource profiles of the team
the data source of the team, e.g. GDRIVE
For teams created from docs, the doc title. Otherwise empty.
when this team was last updated.
whether this team is fully processed or there are still unprocessed operations that'll affect it
Available options:
PROCESSED, QUEUED_FOR_CREATION, QUEUED_FOR_DELETION can this team be deleted. Some manually ingested teams like GCS_CSV or PUSH_API cannot
The logging id of the team used in scrubbed logs, client analytics, and metrics.
A list of objects related to a source object.
The permissions the current viewer has with respect to a particular object.
Unique identifier.
Title or name of the custom entity.
The datasource the custom entity is from.
The type of the entity. Interpretation is specific to each datasource
A list of user roles for the custom entity explicitly granted by the owner.
The permissions the current viewer has with respect to a particular object.
The opaque ID of the Answer.
Example:
3
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The time the answer was created in ISO format (ISO 8601).
The time the answer was last updated in ISO format (ISO 8601).
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The parent board this answer is in.
The collections to which the answer belongs.
The document's document_category(.proto).
Glean Document ID of the Answer. The Glean Document ID is supported for cases where the Answer ID isn't available. If both are available, using the Answer ID is preferred.
Example:
"ANSWERS_answer_3"
Example:
"Why is the sky blue?"
Additional ways of phrasing this question.
The plain text answer to the question.
Example:
"From https://en.wikipedia.org/wiki/Diffuse_sky_radiation, the sky is blue because blue light is more strongly scattered than longer-wavelength light."
The parent board ID of this Answer, or 0 if it's a floating Answer.
Filters which restrict who should see the answer. Values are taken from the corresponding filters in people search.
A list of user roles for the answer added by the owner.
A list of user roles for the answer removed by the owner.
A list of roles for this answer explicitly granted by an owner, editor, or admin.
Available options:
DOCUMENT, ASSISTANT The permissions the current viewer has with respect to a particular object.
Heading text that was matched to produce this result.
Question text that was matched to produce this result.
Example:
{
"snippets": [
{
"snippet": "snippet",
"mimeType": "mimeType"
}
],
"metadata": {
"container": "container",
"createTime": "2000-01-23T04:56:07.000Z",
"datasource": "datasource",
"author": { "name": "name" },
"documentId": "documentId",
"updateTime": "2000-01-23T04:56:07.000Z",
"mimeType": "mimeType",
"objectType": "objectType"
},
"title": "title",
"url": "https://example.com/foo/bar",
"nativeAppUrl": "slack://foo/bar",
"mustIncludeSuggestions": [
{
"missingTerm": "container",
"query": "container"
}
]
}
The attendee list, including their response status
The app or other repository type this represents
The datasource-specific type of the document (e.g. for Jira issues, this is the issue type such as Bug or Feature Request).
Mimetype is used to differentiate between sub applications from a datasource (e.g. Sheets, Docs from Gdrive)
If there is available icon URL.
The unique ID of the Collection.
The unique name of the Collection.
A brief summary of the Collection's contents.
The emoji icon of this Collection.
Indicates whether edits are allowed for everyone or only admins.
The parent of this Collection, or 0 if it's a top-level Collection.
The datasource type this Collection can hold.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The number of items currently in the Collection. Separated from the actual items so we can grab the count without items.
The number of children Collections. Separated from the actual children so we can grab the count without children.
The items in this Collection.
Metadata having what categories this Collection is pinned to and the eligible categories to pin to
The names of the shortcuts (Go Links) that point to this Collection.
The children Collections of this Collection.
A list of user roles for the Collection.
A list of added user roles for the Collection.
A list of removed user roles for the Collection.
Filters which restrict who should see this Collection. Values are taken from the corresponding filters in people search.
The permissions the current viewer has with respect to a particular object.
The unique ID of the Answer Board.
The unique name of the Collection.
A brief summary of the Collection's contents.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The number of items currently in the Answer Board. Separated from the actual items so we can grab the count without items.
A list of user roles for the Answer Board.
A list of added user roles for the Collection.
A list of removed user roles for the Collection.
Filters which restrict who should see this Collection. Values are taken from the corresponding filters in people search.
The permissions the current viewer has with respect to a particular object.
Last file match for a repo
Example:
{
"repoName": "scio",
"fileName": "README.md",
"matches": [
{
"lineNumber": 1,
"content": "Welcome to the beginning",
"ranges": []
},
{
"lineNumber": 2,
"content": "Second line of the file",
"ranges": []
},
{
"lineNumber": 3,
"content": "hello world hello world",
"ranges": [
{ "startindex": 0, "endIndex": 5 },
{ "startIndex": 12, "endIndex": 17 }
]
}
]
}
Link text following go/ prefix as entered by the user.
canonical link text following go/ prefix where hyphen/underscore is removed.
Title for the Go Link
A list of user roles for the Go Link.
The opaque id of the user generated content.
Destination URL for the shortcut.
Glean Document ID for the URL, if known.
A short, plain text blurb to help people understand the intent of the shortcut.
Whether this shortcut is unlisted or not. Unlisted shortcuts are visible to author + admins only.
For variable shortcuts, contains the URL template; note, destinationUrl contains default URL.
A list of user roles added for the Shortcut.
A list of user roles removed for the Shortcut.
The permissions the current viewer has with respect to a particular object.
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The time the shortcut was created in ISO format (ISO 8601).
Example:
{
"name": "George Clooney",
"obfuscatedId": "abc123"
}
The time the shortcut was updated in ISO format (ISO 8601).
Document that corresponds to the destination URL, if applicable.
The URL from which the user is then redirected to the destination URL. Full replacement for https://go/<inputAlias>.
The part of the shortcut preceding the input alias when used for showing shortcuts to users. Should end with "/". e.g. "go/" for native shortcuts.
Indicates whether a shortcut is native or external.
The URL using which the user can access the edit page of the shortcut.
A list of documents related to this structured result.
How this document relates to the including entity.
Available options:
ATTACHMENT, CANONICAL, CASE, CONTACT, CONVERSATION_MESSAGES, EXPERT, FROM, HIGHLIGHT, OPPORTUNITY, RECENT, SOURCE, TICKET, TRANSCRIPT, WITH Which entity in the response that this entity relates to. Relevant when there are multiple entities associated with the response (such as merged customers)
Example:
{
"query": "app:github type:pull author:mortimer",
"label": "Mortimer's PRs",
"datasource": "github"
}
A truncated list of documents with this relation. TO BE DEPRECATED.
A truncated list of documents associated with this relation. To be used in favor of documents because it contains a trackingToken.
The text of the related question
The answer for the related question
Subsections of the answer string to which some special formatting should be applied (eg. bold)
A subsection of a given string to which some special formatting should be applied.
A disambiguation between multiple entities with the same name
Name of the ambiguous entity
The unique id of the entity in the knowledge graph
The type of entity.
Available options:
PERSON, PROJECT, CUSTOMER Any snippets associated to the populated object.
A matching snippet from the document. Query term matches are marked by the unicode characters uE006 and uE007.
The mime type of the snippets, currently either text/plain or text/html.
A matching snippet from the document with no highlights.
Used for sorting based off the snippet's location within all_snippetable_text
The bolded ranges within text.
A subsection of a given string to which some special formatting should be applied.
A URL, generated based on availability, that links to the position of the snippet text or to the nearest header above the snippet text.
An opaque token that represents this particular result in this particular query. To be used for /feedback reporting.
The level of visual distinction that should be given to a result.
Available options:
HERO, PROMOTED, STANDARD Source context for this result. Possible values depend on the result type.
Available options:
EXPERT_DETECTION, ENTITY_NLQ An opaque token that represents this particular result in this particular query. To be used for /feedback reporting.
Response timestamp of the message.
A unique server-side generated ID used to identify a message, automatically populated for any USER authored messages.
Opaque tracking token generated server-side.
Semantically groups content of a certain type. It can be used for purposes such as differential UI treatment. USER authored messages should be of type CONTENT and do not need messageType specified.
Available options:
UPDATE, CONTENT, CONTEXT, DEBUG, DEBUG_EXTERNAL, ERROR, HEADING, WARNING Signals there are additional response fragments incoming.
The id of the associated Chat the messages belong to, if one exists.
Follow-up prompts for the user to potentially use
Time in milliseconds the backend took to respond to the request.
Example:
1100
A token that is used to track the session.
curl --request POST \
--url https://{domain}-be.glean.com/rest/api/v1/chat \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"messages": [
{
"author": "USER",
"messageType": "CONTENT",
"fragments": [
{
"text": "What are the company holidays this year?"
}
]
}
]
}'{
"messages": [
{
"author": "GLEAN_AI",
"messageType": "CONTENT",
"hasMoreFragments": false,
"agentConfig": {
"agent": "DEFAULT",
"mode": "DEFAULT"
},
"fragments": [
{
"text": "There are no holidays!"
}
]
}
]
}