The following page may contain information related to upcoming products, features and functionality. It is important to note that the information presented is for informational purposes only, so please do not rely on the information for purchasing or planning purposes. Just like with all projects, the items mentioned on the page are subject to change or delay, and the development, release, and timing of any products, features or functionality remain at the sole discretion of GitLab Inc.
| Stage | Foundations |
| Maturity | Viable |
| Last reviewed | 2025-09-15 |
Thanks for visiting this direction page on the Design System category in GitLab. This page belongs to the Foundations Stage and Chris Micek maintains it.
This direction page is a work in progress, and everyone can contribute:
The Pajamas Design System (also known as Pajamas or simply the design system) is the foundation for how we design and build product experiences at GitLab. It’s a shared system of design tokens, components, patterns, and guidelines that enables teams across the company to create cohesive, accessible, and high-quality user experiences efficiently and at scale.
Efficiency means solving common design and implementation problems once and making those solutions easy to reuse. Quality means providing the assets, structure, and guidance teams need to create a unified and predictable experience across the entire product.
The Design System group maintains and evolves the design system in close partnership with contributors. We define best practices, improve the quality of components and patterns, and triage issues that span the product. Our goal is to help teams build better experiences faster by defining and strengthening the shared foundation that brings together the work of many teams into a unified product.
Contributing to the design system should feel productive and approachable. In the past, the process has sometimes felt slow, unclear, or overly guarded, which creates friction for teams and limits the system’s ability to grow. In FY26-Q3, we will roll out a new contribution model that clarifies expectations and ownership while making it easier for teams to propose changes. Alongside these process improvements, we will invest in the relationships and support that help contributors work through the model to propose new assets or refine existing ones with confidence.
| Initiative | Description | Timeline | Dependency |
|---|---|---|---|
| Core and extended layers | We’re introducing core and extended layers to the design system. The core layer contains building blocks that are broadly applicable and maintained by the Design System group. The extended layer has lighter governance, offers more flexibility, and teams contribute to and own it. This approach supports more contributions, reduces friction, and reflects the reality of how teams work at GitLab today. | Launch September 15 | None |
| Contribution process and support model | We’re revamping the process that governs how we support contributions. A key part of this change is proactive outreach to all teams with new or active contributions, ensuring every contributor has a clear point of contact to help move their work forward. This includes early reviews, guidance on next steps, and, in some cases, direct support. Alongside this outreach, we’re formalizing the overall process with defined levels of support and a review model that accounts for different types of contributions. | Launch September 15 | None |
To understand the impact of these efforts, we’ll track both perception and participation.
| Indicator | Description | Data |
|---|---|---|
| Pajamas Sentiment | Perception of the design system is just as important as its implementation. Sentiment data helps us understand whether teams find the system effective, reliable, and high-quality. Our last internal survey was in FY26-Q2 and it showed a drop in ease of contribution compared to the FY2-Q2 | Pajamas sentiment survey results Q2FY26 |
| MR Rates for Pajamas | Contribution volume is a key signal of system engagement. Tracking MR activity helps us measure participation and identify trends in contribution over time. | Merge request analytics |
| Number of Unique Contributors | Broader participation strengthens system ownership and resilience. Tracking unique contributors helps us understand adoption and shows whether teams are actively engaging with and supporting the system. | TBD |
| Burndown rate | TBD | TBD |
As GitLab transforms into an AI-native product, Pajamas must provide the foundational infrastructure for new patterns of interaction, layout, and expression. We're actively updating core elements—components, layout structures, and design tokens—to align with this evolved visual direction while establishing comprehensive guidance for AI interaction patterns. By creating these foundational patterns within the design system, we ensure that AI-powered experiences maintain consistency across the platform as teams experiment with new capabilities and use cases.
| Initiative | Description | Timeline | Dependency |
|---|---|---|---|
| Align design system components | Members of the Design System team are collaborating with the Project Studio task force to update and prepare components for the AI-native vision. Changes are driven by rapid iteration and experimentation, not by a centralized update effort. The design system team is focused on supporting this iteration as changes arise and may conduct a broader audit once the direction becomes more stable. | Completed by November to support Duo Agent Platform GA | This has a dependency on a finalized design direction for Duo Agent Platform GA. Roughly half of the Design System team are contributing directly to Project Studio efforts. As this work progresses, we need those involved to identify the changes needed to bring existing components in line with the new design direction. Support will also be needed to perform any migrations in the event of breaking changes. |
| Expanded tokenization | We are expanding the design token system to support new layout style, and interaction patterns required by the AI-native vision. This includes design tokens for spacing, border-radius, elevation, and typography. These additions support a high-quality design in the new panel-based architecture and are foundational to a consistent experience with Duo UI Next. While DUO UI does not use the current design system directly, it can import and integrate tokens, making expanded tokenization critical for long-term alignment. | Completed by November to support Duo Agent Platform GA | The Design System team will lead token implementation across components and maintain the token framework, but support is needed to integrate new tokens into Duo UI, Duo UI Next, and the wider product. As with the dark mode rollout, we'll provide guidance and migration support, but successful adoption will require collaboration from product teams to implement new tokens in their own areas. |
| Design guidance for AI-native interfaces | We will create guidance to help teams adopt the AI-native vision once the design direction has stabilized.. Early focus areas include panel-based architecture and designing within containers instead of traditional responsive breakpoints. Changes are driven by rapid iteration and experimentation, not by a centralized update effort. The design system team is focused on supporting this iteration as changes arise and may conduct a broader audit once the direction becomes more stable. | Completed by November to support Duo Agent Platform GA | This has a dependency on a finalized design direction for Duo Agent Platform GA. This work is owned by the Design System team and does not need additional support. |
| Reduce friction and cognitive load of using the design system | We’re working to reduce friction and the cognitive load of using the design system. This includes making it easier to spin up a project using the design system, easier installation, and making it easier to use for quick prototypes or projects outside the core product. We’re also looking into providing more information about the design system code API for consumption where developers are working, like in AI and IDEs. | TBD | TBD |
| Indicator | Description | Data |
|---|---|---|
| Coverage of AI components within Pajamas | TBD | TBD |
| Availability and coverage of design tokens across the application | TBD | TBD |
GitLab's renewed focus on product-led growth demands design patterns optimized for user activation, engagement, and conversion—critical metrics for achieving the company's growth targets. Through our reimagined contribution model, we're enabling Growth teams to rapidly iterate on conversion-focused patterns and contribute successful experiments back to the system. By facilitating direct collaboration between Growth teams and Pajamas, we ensure that high-performing PLG patterns become immediately available to all teams, accelerating the consistency that drives user engagement and reduces development overhead. PLG-specific guidance will be embedded directly into Pajamas, from onboarding flows that reduce time-to-value to engagement patterns that drive feature adoption.
| Initiative | Description | Timeline | Dependency |
|---|---|---|---|
| Indicator | Description | Data |
|---|---|---|
We're evolving GitLab's visual language to reflect our brand personality within the product experience, moving beyond the dated UI that has remained largely unchanged for years. This work establishes a refreshed design vision with visual examples of updated patterns, layouts, and components, along with guiding principles for implementation across common workflows. Our approach focuses on creating systematic foundations through page templates and reusable patterns that enable rapid, cascading updates instead of requiring changes to hundreds of individual files.
| Initiative | Description | Timeline | Dependency |
|---|---|---|---|
| Modernized design vision | Establish a design vision that shows what a modernized GitLab UI can look and feel like across common workflows. This includes visual examples of updated patterns, layouts, and components, along with a set of principles to guide implementation across the product. The goal is to align on a shared direction that enables a path to implementation. | TBD | |
| Define page templates and Modern UI patterns | To express the updated visual language consistently across the product, we’re identifying common page layouts and reusable patterns that reflect the new design direction. This includes codifying structural templates and high-usage interaction patterns across key workflows. The goal is to provide a foundation for implementing the modernized design vision, helping teams apply the visual language at the page and pattern level. | TBD |
| Indicator | Description | Data |
|---|---|---|
We’re revisiting the current state of components and patterns across the design system. This includes auditing what exists, identifying quality or usability issues, and working through pre-existing feature requests. We use a component health assessment to evaluate where components fall short and where improvements will have the most impact. This work focuses on components and patterns in the core layer of the design system.
| Initiative | Description | Timeline | Dependency |
|---|---|---|---|
| Component health audit | We’re conducting a systematic audit of existing components using the component health assessment framework. Each component is evaluated across four key dimensions: design, code, documentation, and accessibility. The findings will be used to prioritize work in the component maturity epic and to communicate stability and quality to design system consumers. | TBD | None |
| Improve component maturity | Building on the results of the component health audit, we’re improving the maturity of existing components by addressing known gaps and usability issues. This includes fixing accessibility bugs, resolving inconsistencies between design and implementation, updating documentation, and introducing long-requested missing features. The goal is to bring each component to a complete and usable state across the four dimensions of design, code, documentation, and accessibility. | ~6 milestones | None. |
| Improve pattern maturity | In addition to components, we’re focusing on the maturity of design patterns in the core layer of the design system. Many patterns today are incomplete, lacking documentation, or guidance for robust use. This initiative includes identifying high-impact patterns, expanding and refining guidance, and, in some cases, building reusable components to support them. The goal is to better support teams in creating more complex and consistent interfaces across the product | TBD | None. |
To understand the impact of these efforts, we’ll track both perception and progress.
| Indicator | Description | Data |
|---|---|---|
| Pajamas Sentiment | Perception of the design system is just as important as its implementation. Sentiment data helps us understand whether teams find the system effective, reliable, and high-quality. | Pajamas sentiment survey results Q2FY26 |
| Component health | We’ll assess and track the health of core components using defined criteria across design quality, development standards, accessibility, and documentation. | Status Matrix |
We launched dark mode as a fully supported feature across the GitLab product. This was a cross-functional effort that required deep collaboration between the Design System group and teams across GitLab. As part of this work, we transitioned the product to use design tokens for color, replacing hard-coded values with a more flexible system. This shift not only made dark mode possible, but also established a more scalable and maintainable foundation for future theming, visual refinements, and accessibility improvements.
Building and integrating all components across GitLab. The scope of this group is to provide guidance and governance for the design system and related tooling, and is staffed with dedicated product designers and engineers to support that. Creating and implementing components across the product requires participation from every group and category.
Internal product designers, technical writers, engineers, and product managers.