Tabs

Tabs are used to group different but related content, allowing users to navigate views without leaving the page. Tabs can be used on full page layouts or in components such as modals, cards, or side panels.

To help reduce cognitive load, use tabs to group related information into different categories. Tabs can be used to organize content such as forms, settings, and dashboards so a user does not have to navigate away from their workflow to complete their task.

When toggling between different formats of the same content or filtering the same content, use content switcher instead. Content switcher is often used alongside tabs, but it typically serves at a lower hierarchy to organize related content within the tab’s contents.

When the user needs to work through a step by step linear process, use a progress indicator. Tabs help organize information hierarchically but remain flexible in how content can be designed and consumed.

Tabs should not be used if the user needs to compare information in different groups, as this would result in the user having to click back and forth to complete a task.

The tab component consists of two distinct zones: selected and unselected. One tab is always selected by default. There are always at least two tabs in non-dismissible tab groups and at least one in dismissible tab groups. Icons are optional.

A. Label
B. Indicator
E. Container
H. Tablist extended background
G. Tab panel

Much like buttons, alignment of tabs depends on where they appear and whether or not they’re contained within another component. As a general rule, the first label for both line tabs and contained tabs align with the grid and the text below. If tabs are within another component, such as a card, follow the grid that you are using inside the component and align the label with text in the component. Vertical tab labels also align with the grid, so all labels in a tablist will align on the same grid column.

With line tabs and contained tabs, auto-width is the default behavior. Each tab will be a different size depending on the label’s character count but will have consistent padding on each side of the label. The first label, selected by default, should align to the grid. Where the tabs end will vary and may not end on the grid. If needed, you may also use a line to help balance tabs with other components on the page.

Instead of using the default auto-width behavior, contained tabs also have the option to align to the grid. As a group, the tablist spans a set of columns with each tab being equal in size. The first tab’s label should align to the first column you are using with the last tab in the group always ending at a column’s edge. The tabs in between will flow accordingly and may or may not align to the grid but will always be the same size.

Use the 2x grid to drive visual rhythm by spacing content in multiples of two columns and aligning the beginning and ending of the tab elements with content below the tabs when possible.

Use grid aligned contained tabs when:

Depending on breakpoints, vertical tabs align to grid columns spanning 4 or 2 columns, of the 16 column grid and connects to a tab panel that can be 8 or 12 columns wide. See vertical tabs for more information on breakpoints and responsiveness.

When using line tabs within a component (like a modal) the first label should always align to the other content in the space. Do not use contained tabs. The edges of the line tab may also bleed to the edges of the space if needed depending on the container’s margins. To add hierarchical clarity and balance the tabs with the other content, a line may be added from the end of the last tab to the edge of the content area.

Line and tab labels do not need truncation since these types of tabs allow for horizontal scrolling. The tabs themselves can grow and shrink accordingly. However, when a label is too long within a vertical tab, the label overflows to 2 lines and then truncates with an ellipsis. By mouse, the full title is disclosed in a browser tooltip on hover. By keyboard, the full title is disclosed on focus in a tooltip.

For further content guidance, see Carbon’s content guidelines.

Tabs have two main states—selected and unselected. Other interactive states are hover, focus, and disabled. The default view is one tab is preselected and is usually the first tab. Only one tab can be selected at a time. When a user chooses a new item, the previous tab is automatically deselected. If a user navigates away from a tab, it remains the selected tab until altered by the user.

For detailed visual information about the various states for this component, see the Style tab.

When your page requires more tabs that can fit or needs to adapt to a new browser size, line and contained tabs should become scrollable. Left and right arrows appear to help navigate the user through tabs that might be off-page.

Users can trigger a state change by clicking anywhere in the container area.

One tab should be selected by default. Users can navigate between tabs by pressing right or left arrow keys. For additional keyboard interactions, see the Accessibility tab.

Automatic and manual tablists differ in how the tab items are activated.

For automatic tablists, focus and selection are synchronized. When the user arrows to a tab, it becomes selected, and the tab panel under the tab updates automatically. Use automatic tabs when tab panel content can load quickly, allowing for users to scan quickly through information to make a decision without latency.

Manual tablists allow the user to arrow between the tab items without updating the tab panel underneath. When the user navigates using arrows, the selected tab remains selected while focus moves to the next tab. In order to select the focused tab (and update the tabpanel under the tab), the user would press Enter or Space. Use manual tabs when information in the tab panel will take a longer time to load. This will allow a keyboard user or screen reader to navigate through the tablist without having to wait for content to load. Learn more about automatic verses manual tablists in the Keyboard interactions section of the accessibility tab.

For additional information, see the Accessibility tab.

A line tab is standalone tab that can also be nested within components. Although it is commonly used within components, such as a modal or header, line tabs can also be used for large content areas using the entire page for layout, not connected to any other components. Line tabs are highly flexible and can be placed on background or layer tokens, making it easy to use in many different scenarios. To help define the tab panel, which is the content area below the selected tab, a line can be used that extends from the tablist to the end of the panel. This is currently not part of the Carbon line tabs component and must be added by teams if desired.

Contained tabs are an emphasized tab commonly used for defined content areas like sub-pages. Contained tabs are always attached to a tab panel (background container) that uses the same layer token as the selected tab. Since the contained tabs are on layers, the tab content tends to stand out against the background and maintain a high visual hierarchy while line tabs tend to blend into content more easily. Because of the layering model, be mindful of the layers used within the tabs so the content does not get too visually complicated, especially within smaller areas.

Vertical tabs are left, vertically aligned tabs that allow a user to scan through information from top to bottom. This vertical orientation is good for quickly browsing and accessing information, such as a [get started] (https://pages.github.ibm.com/cdai-design/pal/patterns/onboarding/orientation/usage#get-started) pattern. Do not use vertical tabs in place of navigation.

The tab panel should stay the same height as tabs switch so content outside the tabs stays in a consistent place. The vertical tablist and panel height should be determined by the tab that has the most content. To avoid excessive scrolling within the tab panel, do not overload tab panel with too much content. If more space is needed than the tab panel allows, consider using line or contained tabs that allow for full page layouts. Both the tablist and the tab panels are always on the same layer. At extra large and large breakpoints, vertical tablist spans 4 columns of the 16 column grid. At medium breakpoint, the vertical tablist spans 2 columns. At the small breakpoint, the vertical tablist uses scrolling contained tabs.

You may use icons in either line or contained tabs. Icons are always pinned to the right of the tab and do not appear above, below, or to the left of the label.

You may use icon-only tabs with both line and contained tabs. Icons must be easily recognized and globally accepted. These work best in small, defined spaces and in components. Always use a tooltip for an icon description on hover and focus to add clarity.

Contained tabs that align with the grid allow for a secondary label. Do not use secondary labels with line tabs, auto-width contained tabs, or vertical tabs.

Dismissible tabs allow users to close tabs providing a focused and relevant experience. Users can add or remove contained or line tabs as needed, accommodating future content additions or modifications without drastically changing the overall layout or structure.

Carbon provides an option to include icons within the dismissible tabs. Only use icons within dismissible tabs if all tabs will include icons. Do not mix dismisible tabs without icons with dismissible tabs with icons.

When dismissing a tab, an inline warning or modal may be used when information contained in a tab will no longer be accessible or difficult to retrieve. Since warnings, especially modals, are highly disruptive, only use when the dismissal causes errors, unintentional deletions, or unsaved changes.

There are various ways to trigger a new tab. The trigger button can visually change its shape and size depending on the use case. Keep the trigger close enough to the new tab to associate the add action with the new tab item. The order of tabs can be ascending or descending depending on use case but do keep them in a sequential, logical order.

If all tabs are dismissible, make sure a user understands how to trigger new tabs once the tabs are gone. Provide visual cues, such as a container or placeholder tab, so the user clearly understands that the trigger button is creating a new tab in a place they expect.

Content switchers allow users to compare and toggle between alternate views of similar or related content. Content that is grouped into tabs is part of the same bigger context but the content does not overlap.

Progress indicator content moves in a logical progression, showing next steps to guide the user through the completion of a task. Tabs organize content into groups that the user can navigate through; they do not support progressive tasks.

Jakob Nielson, Tabs, Used Right (Nielsen Norman Group, 2016)

Help us improve this component by providing feedback, asking questions, and leaving any other comments on GitHub.