Changelog

We've improved the anomaly detection experience with better cell highlighting in Retention Tables and the ability to toggle anomaly visibility on reports.

Users working with AI assistants couldn't easily find and reference their scheduled reports through MCP. Now, you can search for schedules via MCP, enabling AI assistants to help you find, reference, and manage your scheduled analytics workflow.

When working with MCP, AI clients didn't know when to automatically pull raw data vs. just show summaries, leading to suboptimal responses. Now the MCP server provides intelligent hints based on data size and number of slices, helping AI assistants automatically determine when to fetch detailed data for better analysis.

The "Authorize Client" page couldn't reliably identify which AI client was connecting, showing only redirect URLs. Now users see friendly names for known providers like Claude and Cursor during authorization, making the connection process clearer and more trustworthy.

AI Detect was unnecessarily running anomaly detection on reports with fixed date ranges, which doesn't make sense since the data doesn't refresh. Now, reports with fixed date ranges are intelligently skipped, reducing noise and focusing AI Detect on reports where anomaly detection is meaningful.

You can now configure notification preferences per Dashboard when adding a Slack channel, giving you granular control over which alerts go to which channels.

When creating or editing a report, some of the graphics and visual elements weren't optimized for dark mode, reducing readability. Now, improved color adjustments in dark mode provide better visibility and a more polished experience for users who prefer dark themes.

Kubit's MCP makes it incredibly easy to work with your data directly inside AI tools like Cursor and Claude. Instead of switching between platforms, you can simply ask to explore your data model, create reports, review results, export underlying data, or search for existing reports and dashboards - all from within your AI workspace.

We moved the Report builder to the left-hand side to create a more efficient and comfortable workflow. This layout lets users view the builder and the report side by side, reducing scrolling fatigue and speeding up interactions.

Kubit's Flow Analysis is now fully supported on ClickHouse. Teams using ClickHouse can now explore user journeys, identify drop-offs, and understand conversion patterns with the same powerful flow visualization capabilities.