New Features

Cloud-Native Deployment with Workspace Management

Flexible deployment model that works both locally and in cloud environments:

• Location-Independent Deployment - Server can now start anywhere without requiring a fixed project folder using the --cloud flag
• GitHub Project Sync - Create projects and sync back from GitHub repositories for centralized management
• Workspace Organization - Organize multiple projects, configurations, and resources within workspaces
• Environment-Based Configuration - Read database connection info and addresses directly from environment variables
• Backward Compatible - Traditional folder-based deployment model remains fully supported

The new cloud-native mode enables flexible deployment scenarios while maintaining support for the original local development workflow.

Agentic Workflows

Transform AI-driven data analysis into reliable, production-ready systems with structured workflow primitives. Unlike traditional predefined task sequences, agentic workflows enable agents to autonomously plan and execute multi-step analytical processes, bringing software development lifecycle principles (build-test-deploy) to AI-powered data workflows.

Synchronous APIs

New API endpoints for non-streaming operations:

• Workflow Sync APIs - Synchronous endpoints for workflow operations that previously required streaming
• Complete Response Model - Get full responses immediately for integration scenarios requiring non-streaming behavior

Platform Improvements

Enhanced stability and compatibility across the platform with improved semantic layer engine for better data modeling, ClickHouse row filtering API, Domo DECIMAL type support and sync improvements, various IDE bug fixes, and updated dependencies to latest versions.