Release Notes Incorta v8.2 - 2026.8
Release date: October 12, 2026
Release highlights
This release delivers enhancements across AI, data lake interoperability, Data Catalog, analytics, and security, including:
- Build Incorta Applications: Incorta AI Applications are now generally available. Describe an interactive application in plain language and have it running in minutes on the same governed data and permissions you already trust; or develop locally with AppKit.
- Connect your AI with Incorta: The new Incorta MCP Server (preview) lets AI clients such as Claude securely explore and query Incorta data, with no custom integration code.
- Iceberg joins Delta as a first-class external table: Load Iceberg directly into memory from standard REST catalogs without creating a copy of data.
- Govern at a column level: Document, assign, search, and audit tags down to individual columns, and carry them between environments with Data Catalog exports.
- Compare patterns, not just numbers: Trellis charts turn one insight into a grid of aligned small charts, one per segment, revealing differences in trend and shape that a single combined chart hides.
- Security and Access: Single active-session enforcement across browsers and devices, and a new Public API endpoint for securely querying accessible business views.
Upgrade considerations
Upgrade considerations for previous releases also apply to this release unless stated otherwise.
Data agent supported version
This release supports Data Agent 16.2.0. The data agent upgrade instructions for 2026.3.0 still apply.
Behavior Changes
Incorta reserved columns
Encryption is now disabled for the Incorta reserved column reserved_incorta_extraction_timestamp, preventing extraction failures when writing Parquet files.
Global variable accessibility for all users within the tenant
Global variables now appear for every user in the formula builder and in the presentation variable list. After upgrading, all users in a tenant can see and use every global variable.
Previously, non-admin users could only use the global variables they had created themselves. For example, an Analyze user without access to the Global Variables page could not find a global variable created by another user in the formula builder. Access to the Global Variables page itself is unchanged.
This matches the documented behavior that global variables are not subject to user access rights and are available to all users in the tenant. For more information, refer to Concepts → Global Variable.
Some contexts do not support global variables. For the list, refer to Unsupported global variable usage.
String presentation variables now require existing variables
When you set a String presentation variable to a variable, you must now select the variable from the dropdown list. Previously, you could type any name starting with $ (for example, $myVariable), even if no variable with that name existed. Incorta saved the typed text as a plain string instead of a variable, so no error appeared.
New features
Incorta AI
Data Management
- External and remote table support for Apache Iceberg tables
- Read external Delta tables through Unity Catalog with credential vending
- Advanced SQLi support for Delta and Iceberg external tables
Data Governance
Dashboards, Visualizations, and Analytics
- New Trellis charts
- Conditional formatting for Aggregated and Pivot tables subtotals
- Abbreviation option when viewing Charts as Tables
Architecture and Application Layer
Incorta AI
Incorta Applications generally available
This release introduces the first generally available (GA) version of Incorta Applications. Applications turn the same governed Incorta data foundation into purpose-built, interactive applications you can describe in plain language and have running in minutes. Unlike dashboards, applications provide dedicated pages, forms, databases, and data storage to support data entry, workflows, and task management alongside analytics. Applications can read Incorta data in real time, using the signed-in user's permissions, and store application-specific data in their own databases.
Users build and refine applications using the Incorta Builder Agent through natural-language prompts and conversations. Alternatively, developers can use AppKit, a command-line interface (CLI) that supports local development with their preferred editors and tools, and push application versions back to the Incorta Platform. Both tools work with the same applications and version history.
Users must be assigned the new Application Manager role to create and manage applications. All users with access to the Content Manager can view and use applications from the Dashboard Directory. However, when sharing an application with users, you must also share view access to the underlying business schema or schema.
The new Incorta Applications GA feature is available on Incorta Cloud deployments only, and is a metered, pay-as-you-go capability. Compute, memory, and data transfer are tracked per application and per tenant, and billed based on actual usage. The feature is turned off by default. To enable it, contact your Incorta Account Executive.
- Applications can read Incorta data but cannot write to Incorta.
- All application data is stored in the application's own database.
- Applications respect existing access permissions and row-level security.
- Queries execute as the signed-in user, subject to the same access controls as other Incorta queries.
After upgrading to this release, the AI Apps tab will not be available, and applications created using the AI Apps preview release introduced in v8.0.0 will not be automatically migrated. However, you can still access these applications at <cluster_url>/data-apps, for example, https://my-cluster.cloud.incorta.com/data-apps. Use your Incorta login credentials to sign in.
Incorta MCP Server (Preview)
Incorta now provides an MCP Server for Incorta Cloud, enabling AI clients, such as Claude, to securely access Incorta data without custom integration code.
Each cluster provides its own MCP server. Users connect via OAuth. The server does not store credentials; it forwards them to Incorta.
This initial version of the Incorta MCP server includes only one toolset, query_data, which allows an AI client to list the schemas a user can access, inspect tables and columns, and run read-only Spark SQL queries against them.
The feature is turned off by default. To enable it, turn on the Incorta MCP toggle in the Cloud Admin Portal to get the MCP URL, which is used to configure your AI client.
For more details, refer to References → Incorta MCP Server and Guides → Integrate Incorta MCP Server with Claude.
Data Management
External and remote table support for Apache Iceberg tables
In addition to supporting Delta tables as external and remote tables, Incorta now supports reading Apache Iceberg tables through the new Iceberg REST catalog connector or the Data Lake connectors. The Iceberg REST catalog connector supports multiple Iceberg Catalogs, such as Polaris, Nessie, and Lakekeeper. Incorta Data Lake connectors can also be used to discover Iceberg tables stored in cloud storage (GCP, ADLS, and S3). You must upgrade the connector to an Iceberg-compatible version.
Incorta does not support Iceberg REST catalogs that use AWS Signature Version 4 (SigV4) authentication, such as AWS Glue and Amazon S3 Tables.
Read external Delta tables through Unity Catalog with credential vending
Incorta can now discover and read external Delta tables managed by Databricks through the Unity Catalog (through the new Databricks Unity Catalog connector). Unity Catalog provides temporary credentials for the underlying storage (credential vending). You no longer need to provide credentials for the underlying Databricks external storage. Unity Catalog issues temporary, table-scoped credentials when reading the table data. Additionally, schemas and tables appear with their Unity Catalog user-friendly names instead of the storage-level GUIDs used before.
Advanced SQLi support for Delta and Iceberg external tables
Previously, Advanced SQLi treated external Delta tables as remote tables. Starting with this release, Advanced SQLi now fully supports Delta and Iceberg external tables, including their joins and formula columns. This ensures that queries from Spark SQL Views and external tools connected through Advanced SQLi, such as Tableau and DbVisualizer, return results consistent with Incorta insights.
Data Governance
Support for documentation and column-level tags
Tags now support column-level assignments, documentation, and searching and filtering columns by tags. Tag operations, including creation, changes, assignments, and documentation updates, are audited. Data Catalog exports and imports include tags’ definition as well as documentation and column assignments. The new file tags.csv holds tag definitions and their documentation, while assets_tags.csv holds tag assignments, including column-level ones. Data Governors add documentation to tags, while users with schema edit permissions can assign tags to columns.
Data Catalog Documentation editor enhancements
The Documentation editor now shows a character count and the percentage of the available size used. The indicators update as you type. Once you exceed the size limit, you cannot save the documentation.
The size limit is also enforced during import. Documentation that exceeds the limit is not imported.
This enhancement is available in the Documentation section for data assets, domains, glossary terms, and tags within the Data Catalog.
The following count toward the size limit:
- Text
- Formatting, including bold, colors, and font changes
- Non-Latin characters, which use more space than Latin characters
Dashboards, Visualizations, and Analytics
New Trellis charts
You can now display a chart as a trellis: a grid of small charts, one for each value of a dimension, by using the Break By tray when configuring an insight in the Analyzer. You can define the number of charts per row and column and choose whether to unify the X-axis, Y-axis, and Categories, enabling you to compare subsets of your data side by side without creating and aligning separate insights. Trellis is supported for most chart types with an X-axis and Y-axis.
Avoid using dimensions with a large number of values, as the resulting number of charts may be too large to render.
Conditional formatting for Aggregated and Pivot tables subtotals
This release introduces conditional formatting to subtotal values in aggregated and pivot tables. When Totals and Subtotals are enabled under Settings > General, you can apply conditional formatting by navigating to Measures > Pill Properties > Appearance and using the new Subtotal Cell Styles and Grand Total Cell Styles options. You can set conditional formatting rules, such as displaying values greater than 3000 with a specific text or cell background color.
Abbreviation option when viewing Charts as Tables
In this release, when viewing a chart as a table by selecting More Options > View as Regular Table or View as Aggregated Table, you can now control whether values are abbreviated. The abbreviation option uses each measure’s abbreviation setting from the Analyzer. Measures configured with None remain displayed as full values.
Architecture and Application Layer
Single active session enforcement
This release introduces a new CMC option to restrict each user to a single active session across devices and browsers. When enabled, a new sign-in automatically invalidates the user's existing session across all authentication methods, including Incorta login and SSO, and all nodes in multi-node deployments.
To enable this feature, in the CMC, go to Server Configurations > Security, turn on Enable Sessions Tracking and Control and Allow Only One Active Session per User, and then restart all the Analytics services.
New business view query endpoint
This release introduces a new Public API endpoint, /{tenant}/query/business-views/, to query only business views that the user owns or that are shared with them. If the user does not have access, the query is not run and no data is returned.
The new endpoint works the same as the existing /{tenant}/query endpoint, including request and response formats, pagination, limits, and security filters. The only differences are the access check for the specified business view and the URL path.
The existing /{tenant}/query endpoint continues to work, but it will be retired in a future release. We strongly recommend moving your integrations to the new endpoint to ensure business view access rights are respected. Because the new endpoint returns data only for business views the user has access to, users and service accounts running your integrations must have access to the business views they query.
Enhancements
| Description | Area |
|---|---|
| Incorta has upgraded Quartz to the latest version 2.5.2. | Scheduler |