Getting started: Release notes
Changelog for Linkurious Enterprise v4.3.4
Release Date: 2026-04-08
New Features & Improvements (28)
- (v4.3.3) CSV Data Import: Improve user experience
- (v4.3.3) Data Source: Improve the data source connection user experience
- (v4.3.3) Google BigQuery: Adding support for BigQuery Graph
- (v4.3.3) Query AI: Added support for new LLM vendors (Azure OpenAI, Amazon Bedrock), and added optional custom URLs (Mistral AI, Google)
- (v4.3.3) Visualization: Context menu user experience improvements
- (v4.3.3) Visualization: Open a graph selection in a new Visualization (with copy-paste or right click)
- (v4.3.3) Google Spanner Graph: Add an option to add a custom query hint to all queries
- (v4.3.3) Query Builder: Remove "NEW" tag
- (v4.3.3) Security: Stop using the "axios" library
- (v4.3.3) Security: Update several internal dependencies
- (v4.3.2) Data Import: New feature to import graph data from CSV files
- (v4.3.2) Alternative IDs migration script: New script to migrate to alternative IDs without loosing visualizations and alert cases
- (v4.3.0) Query AI: The feature is generally available
- (v4.3.0) Visualization: Improve the design of the property panel (closed by default, resizable, show long text values, fixed headers in table view)
- (v4.3.0) Authentication: Add a default value for
access.loginTimeout - (v4.3.0) Browser Support: Updating the list of supported browsers
- (v4.3.0) Deployment: Support defining an application-wide HTTP proxy using environment variables
- (v4.3.0) Elasticsearch: remove official support for Elasticsearch 7
- (v4.3.0) Embedded Elasticsearch: Remove the Embedded Elasticsearch on macOS
- (v4.3.0) Embedded Elasticsearch: Update to Elasticsearch 9.2
- (v4.3.0) Image Export: Deprecate the plugin-in in favor of the application feature
- (v4.3.0) Neo4j: Drop technical support for Neo4j 4.3 and older
- (v4.3.0) Queries: Improve the timeout behavior with workers when some queries run for a very long time
- (v4.3.0) REST Client: Change the constructor parameters of the REST client (agent)
- (v4.3.0) Schema: Remove the default visibility option from the start sampling API
- (v4.3.0) Search API: Allow more than one value for
categoriesOrTypeparameter - (v4.3.0) Security: Increase the default password complexity requirements for local accounts
- (v4.3.0) Security: Update to Node.js v24
Bug Fixes (60)
- (v4.3.4) OpenIDConnect: Login fails when attemting to load user-info URL
- (v4.3.3) Authentication: Wrong screen is shown when starting linkurious for the first time
- (v4.3.3) AWS Neptune: Impossible to add a data-source via AssumedRole ARN
- (v4.3.3) Datasource management: Switching From neo4j / ES to Neptune lead to incoherent index server form
- (v4.3.3) Design Panel: Styles and Captions tabs are broken
- (v4.3.3) Documentation: Missing node grouping default collapse option
- (v4.3.3) Infrastructure: The HTTP_PROXY environment variable is not taken into account for external HTTP queries
- (v4.3.3) Memgraph: Data model feature is missing some details
- (v4.3.3) Node grouping: Error when activating Node Grouping on browsers where GPU is not available
- (v4.3.3) Query AI: Saved query not appearing in the queries panel
- (v4.3.3) Query Template: malformed default value for Date parameters
- (v4.3.3) Upgrade: New scripts are not automatically deployed
- (v4.3.3) Visualization: missing button to manage custom actions
- (v4.3.3) Visualization: Toggle buttons are missing from the Filters section (right panel) for older Chrome versions
- (v4.3.2) Azure Cosmos DB: Editing an existing query causes an error
- (v4.3.2) Custom Actions: Incorrect text labels in the Custom Actions edit modal
- (v4.3.2) Design Panel: Reset Styles pop up should have Close button instead of Cancel button
- (v4.3.2) Documentation: wrong specs for Entity Resolution APIs
- (v4.3.2) Google Spanner Graph: In alerts, some valid case attribute queries fail to run
- (v4.3.2) Neo4j Block Storage: The migration script does not migrate the case visualizations
- (v4.3.2) Neo4j Block Storage: The migration script does not work on Neo4j v4.4.8
- (v4.3.2) Neo4j Block Storage: The migration script doesn't scale past 2.1B nodes/edges
- (v4.3.2) Neo4j Block Storage: The migration script fails to start when the DB version does not match the current version
- (v4.3.2) Neo4j Block Storage: The migration script sometimes fails to find the Neo4j database
- (v4.3.2) Neo4jSearch: Infinite progress percentage during indexing
- (v4.3.2) Neo4jSearch: the reported indexing progress is inaccurate
- (v4.3.2) Node grouping: When enabling node grouping the layout always switches to force-directed
- (v4.3.2) Query Builder: Filter edges by date fails with an error
- (v4.3.2) Schema: The option to enable "View & Search" for a property is available on a node category that is set to "View Only"
- (v4.3.2) Unified Case List: Filtering cases by date shows an error
- (v4.3.2) Visualization: After running a query, the result is sometimes not shown in the visualization
- (v4.3.2) Visualization: Populating of visualization by query ID (deep link) fails to run the query for any data-source other than the first one
- (v4.3.1) Queries: The Cypher auto-complete menu is sometimes misplaced
- (v4.3.0) Alerts: An error is shown in the license modal when alerts are disabled
- (v4.3.0) Alerts: Hiding columns in the Unified Case List only takes effect after refreshing the page
- (v4.3.0) Alerts: In the unified case list, the model filter disappears once applied
- (v4.3.0) Azure SSO: After upgrading to the v2.0 endpoint, we sometimes get the wrong email for users
- (v4.3.0) Data model: Missing styles when you open the data model from the dashboard
- (v4.3.0) Data-sources: Information in the data-sources admin page is outdated
- (v4.3.0) Data-sources: Sometimes, when a data-source cannot connect, the whole application does not load
- (v4.3.0) Deployment: The application is not starting when installed as a service on Red Hat 9.4
- (v4.3.0) Documentation: Misleading "Shortest path analysis" in the feature map
- (v4.3.0) Documentation: Missing timeline zoom limitations
- (v4.3.0) Documentation: Specify that the "id" field is required for Azure Search
- (v4.3.0) Entity Resolution: Missing confirmation pop up when cancelling after deleting a mapping attribute
- (v4.3.0) Google Spanner Graph: It is not possible to configure multiple data-sources for different graphs in the same database
- (v4.3.0) Node grouping: An error is shown when deleting a mix of regular and grouped nodes
- (v4.3.0) Node grouping: Node counters are not recomputed after an "undo" action
- (v4.3.0) Node grouping: Node groups are not always collapsed by default
- (v4.3.0) Node grouping: When selecting a node group, filtered-out nodes are displayed in the table
- (v4.3.0) Queries: For node inputs, the autocomplete fails if the node input specifies more than one node label
- (v4.3.0) Search: category filter in query preview is not applied
- (v4.3.0) Search: In some places, the application is using a hardcoded fuzziness instead of the configured defaultFuzziness
- (v4.3.0) Search: Remote items are sometimes returned when local search is selected
- (v4.3.0) Telemetry: Possible file handle leak when synching archived events
- (v4.3.0) Visualization: After pinning and expanding a node group, the pin appear larger
- (v4.3.0) Visualization: In the property panel, IDs are not aligned correctly
- (v4.3.0) Visualization: In the selective expand modal, the "limit" input is too small
- (v4.3.0) Visualization: Thumbnails are sometimes very different from the actual visualization
- (v4.3.0) Visualization: Wrong values in the timeline after removing items from the visualization
Breaking changes (10)
- (v4.3.0) The default session timeout is now 3 hours: users get logged out after 3 hours of inactivity. When the
access.loginTimeoutconfiguration is not set, the default value is now 10800 seconds. The previous behavior was to never expire sessions when this configuration was not set. - (v4.3.0) We are removing support for browser versions released more than 2 years ago. The oldest version supported is now 119 for Google Chrome, 119 for Microsoft Edge, and 120 for Firefox.
- (v4.3.0) If you are using environment variables like
HTTP_PROXY,HTTPS_PROXYorNO_PROXYto enable proxying of outgoing HTTP traffic, you must now setNODE_USE_ENV_PROXY=1for these environment variables to be taken into account. If you were using the "proxy" option for a Neo4j data-source, this option is now ignored and you must now use the aforementioned environment variables. - (v4.3.0) Deployments that use Elasticsearch 7 must update to Elasticsearch 8+
- (v4.3.0) If you are running the Linkurious Enterprise server on macOS, Embedded Elasticsearch won't be available on macOS anymore. If you need an Elasticsearch instance, we advise you use another way to host Elasticsearch. Please refer to the official Elasticsearch documentation to host it yourself on macOS or other platforms.
- (v4.3.0) The Image export is no longer an official plug-in, the same functionalities are now integrated in Linkurious Enterprise.
- (v4.3.0) The lowest supported Neo4j version is v4.4.0.
- (v4.3.0) The optional parameter
agentof theRestClientconstructor has been removed. This parameter was normally not meant to be used, besides a few advanced specific use-cases. - (v4.3.0) The POST /api/admin/:sourceKey/schema/sampling/start API endpoint does no longer accept a defaultVisibility parameter
- (v4.3.0) When using default settings, password for local accounts must now have at least 12 characters.