🎉 New Features

Schemas

Field-Based Relationships

Schema fields can now be of type relationship, referencing entities from a target schema with one or many cardinality.

Relationship fields support:

• Configurable display templates using variable interpolation for rendering referenced entity labels
• Display modes: list (default) or table for many-cardinality relationships
• Validation rules (required)
• Entity relationship records with schema-path indexing for efficient lookup
• Managed relationship lifecycle - relationships are automatically created and removed when entity fields change

Convert to Relationship Field (Beta)

Existing schema relationships can be converted to inline relationship fields. This is currently in Beta testing - reach out if you want to transfer some of your historical Relationships into inline relationship fields.

Timezone-Aware Datetime Fields

Date fields support two new configuration options:

• Default Format - controls the date/time display format (e.g., date-only, short date-time, long date-time).
• Timezone Display - controls whether timezone information is shown alongside date values.

A User admin settings page, User Settings > Timezone & Locale, provides system-wide defaults.

Experiences

Batch Entity Context

A new batch entity context type collects descendant entity contexts for coordinated batch save operations and unified dirty-state tracking. This enables multi-entity forms where all changes are saved together.

CRUD Button Actions

Buttons now support a CRUD action kind with four operations:

• create - create a new entity from the bound context.
• update - save changes to an existing entity.
• delete - Remove the bound entity.
• save - save all dirty entity contexts (works with batch contexts).

On-success behaviour is configurable: reset the form, refetch data, or navigate to another experience.

This means that you no longer need to create a Flow to save Entities!

Entity Context Field Overrides

Entity contexts support field overrides with two modes:

• default - seeds the value on new entities; remains user-editable
• enforce - always overrides the value at save time; renders bound inputs as read-only

Override values are resolved via data expressions, enabling dynamic defaults from URL parameters, parent contexts, or other datum references.

Quality of Life Improvements:

• Query-Based Entity Context Loading. Entity contexts can now load entities by query with filter and ordering configuration, enabling data-driven experience rendering without requiring explicit entity IDs.
• Entity Context UI Improvements. Entity context configuration in the experience editor has been migrated from switch-based controls to select-based controls for better usability. Schema reference handling has been improved.

Flows

JavaScript Mode for Data Expressions

The data expression node now supports a js mode backed by QuickJS, alongside the existing expression mode. JavaScript mode enables:

• Full JavaScript syntax for complex transformations
• Temporal API support (Instant, ZonedDateTime, PlainDate, PlainTime, PlainDateTime) with arithmetic and comparison operations
• String and vector similarity functions

SQL Integration Improvements

SQL credentials and query workflows received several improvements:

• Connection testing with provider-specific validation (PostgreSQL and SQL Server).
• Connection failure management with automatic state tracking and recovery.
• Connection pool lifecycle improvements (proper close/cleanup on failure).
• UI refinements for SQL credential forms and connection status display.

Query Node Result Limits

The object:query The flow node now enforces a maximum result limit of 1000 records per execution. The entity listing API validates limit at the input boundary.

AI Model Updates

• Added GPT 5.4 and GPT 5.5 models to OpenAI provider.
• Added Opus 4.7 to Anthropic provider.
• Deprecated Claude Haiku 3, Haiku 3.5, Sonnet 4, and Opus 4.

AI Data Extraction Standard Fields

AI data extraction mapping now includes standard fields (e.g., createdAt, updatedAt) alongside schema-defined fields, enabling richer extraction mapping configurations.

Text Template Variable Formatting

Fixed text template variable formatting options not being applied during flow execution. Format directives in content serialisation are now correctly resolved.

List Filter Distinct By Multiple Fields

The List Filter node's "Distinct By Field" option now supports selecting multiple fields for distinctness checks.

Spaces

Space Search Relevance Sorting

Space dashboard search results are now sorted by relevance rather than creation order.

Fixes

Schemas

• Fixed managedRelationships injection during entity updates - relationships are now correctly synced on update, not just create.

Experiences

• Fixed nested ForEach loop datum path resolution when iterating over different datum iterations.
• Fixed Dynamic Link Block URL input visibility when unfocused, variable display, and nested field rendering.
• Fixed entity context settings switching from switches to selects for consistency and corrected schema reference handling.

Flows

• Fixed Temporal values marshalling in data expressions - Instant and ZonedDateTime objects are correctly serialised in Data Expression.
• Fixed Instant and ZonedDateTime arithmetic and comparison operations in Data Expression.
• Fixed date coercion in expression evaluations - mixed date/temporal operands are now correctly coerced.
• Fixed text template variable formatting options not being applied during flow execution.
• Fixed SQL connection pool close warning on connection failure.

Spaces

• Fixed datetime timezone conversion not being applied to table view rendering.
• Fixed timezone and offset mismatch in date display components.