Web analysis without Google: Umami 3.1 brings replays and custom dashboards

Contents

With version 3.1.0, the open-source web analysis tool Umami receives several major new features for analysis, debugging, and performance monitoring. The focus is on freely configurable dashboards called “Boards” and session replays, which allow actual user sessions to be traced. In addition, it includes integrated tracking of Core Web Vitals, significantly expanded filter and segmentation functions, and revised sharing options. The release also closes several security vulnerabilities, performs schema migrations, and raises the minimum platform requirements.

Umami is a privacy-friendly, self-hostable alternative to classic web analytics services like Google Analytics. The software records page views and events without third-party tracking and is primarily aimed at developers and companies who want to control their analysis.

The new Boards offer a flexible dashboard system. Users can combine analyses from charts, tables, and key figures in a freely designable grid. Individual components can be assigned to different websites, adjusted live, and then shared or duplicated within a team. This brings Umami functionally closer to specialized monitoring and BI tools. For example, a marketing team can maintain a board for campaign key figures, while the product team simultaneously manages one for feature usage and conversions.

Session Replay is also new. The function replays real user sessions in the browser and is based on the rrweb library. Session Replay records clicks, scrolling movements, and input. Sensitive content can be hidden via configurable masking levels. Recordings are segmented per visit and can be filtered by events. This helps, for example, to analyze drop-offs in forms or checkout processes.

Performance monitoring and more refined analyses

In the performance area, Umami now captures the Core Web Vitals directly in the visitor's browser, including Largest Contentful Paint (LCP), Interaction to Next Paint (INP), Cumulative Layout Shift (CLS), First Contentful Paint (FCP), and Time to First Byte (TTFB). A revised performance view displays the values and rates them based on common thresholds. This allows, for example, to see if blocking JavaScript is delaying interactions on mobile devices.

The filter and segment logic also grows noticeably. New features include OR conditions between criteria, regex operators, and multiple selections for comparison operators. UTM parameters are now available system-wide. This allows for more complex analyses to be formulated, such as user segments across multiple campaigns or regions. Funnels also support filters on event properties per step and wildcards in goal definitions.

The revised share function allows for finer control of shared content. Individual areas such as overviews or events can be shown or hidden, and share links can be given a name. The display is fully optimized for mobile devices.

Administration, internationalization, and minor improvements

Further innovations include the additional filter dimension “Distinct ID”, flexible time aggregation by hour, day, or month, a download for reports, and performance optimizations through pagination limits and cache-control headers. Additionally, there are geolocation headers for EdgeOne and further configuration options via environment variables.

In the administration area, Umami switches from react-intl to next-intl and provides complete translations for 51 languages. With “react-zen,” the project also introduces a uniform design system and redesigns the navigation. Team functions are now consolidated and supplemented by Redis-based feature control.

Security, migrations, and bug fixes

On the security side, the release closes IDOR vulnerabilities in reports and segments and restricts share tokens more strictly. Additionally, it updates several dependencies with known vulnerabilities.

For the new features, Umami automatically performs schema migrations, for example, for Boards, Session Replay, and the sharing mechanisms. The minimum requirement increases to Node.js 22 because the release relies on Prisma 7.

In addition, version 3.1.0 fixes numerous bugs, particularly with PostgreSQL compatibility, in SQL queries, time zone processing, and in the tracker and user interface. Details on all changes can be found in the Release Notes on Umami's GitHub project page. Version 3.0 was released in November 2025 and added new tracking features to Umami.

(fo)