Sonner Component Guide

Quick Answer

The Sonner component provides the toast viewport and notification behavior used for transient system messages. Use it when a product team needs support for tasks such as confirming completed actions, showing non-blocking errors, and announcing background status. It should feel predictable for designers, engineers, keyboard users, touch users, and people reading long English or Tamil labels.

What Sonner Solves

Sonner exists to communicate system status at the right level of interruption. In a real product, this component is not just a visual style. It defines how teams talk about behavior, state, spacing, and responsibility in review.

For Tamil Design System, the content problem matters as much as the component contract. Labels may be English, Tamil, or bilingual. Text can be longer than the demo version. Users may be on mobile devices, zoomed browsers, or public-service workflows where mistakes are costly. A good Sonner implementation keeps those realities visible.

The practical value is consistency. Designers get a known pattern. Engineers get a maintained primitive. Reviewers get a checklist for severity, timing, dismissal, live-region behavior, focus movement, and recovery paths. Users get an interface that behaves the same way across forms, dashboards, navigation, and tools.

When to Use Sonner

• Feedback is short and temporary.
• The user can continue working.
• The message does not need persistent reference.

Good use of Sonner starts with intent. The component should answer a real product question, not merely fill space. When the purpose is clear, the visual treatment, copy, and interaction states become easier to review.

When to Avoid Sonner

• The content is required to finish a task.
• The error belongs beside a form field.
• The message needs long explanation.

Avoiding the component is sometimes the more mature design decision. If the task needs a different semantic control, a full page, a simpler text explanation, or a more explicit confirmation, choose that pattern instead of forcing Sonner to do too much.

Anatomy and Behavior

• Toast viewport.
• Toast title.
• Description.
• Action.
• Dismiss control.
• Duration.

Each part of Sonner should have a job. The root controls structure, the visible label communicates purpose, and the state treatment tells users what is possible now. If one of those responsibilities is missing, the component may still look correct in a screenshot but fail in production.

Behavior should be boring in the best way. Hover, focus, active, selected, disabled, loading, empty, and error states should follow the same interaction language used across Tamil Design System. That consistency helps users predict what will happen before they interact.

Props and Implementation Contract

• Position controls viewport placement.
• RichColors can map severity.
• Duration controls visibility.
• Action supports quick recovery.

Treat props as a product contract, not as a dumping ground for visual switches. A useful prop changes behavior, semantics, state, or a clearly named visual hierarchy. If teams need custom styling, allow scoped class extension, but keep the default component complete enough to use without local patching.

Implementation should also preserve native browser behavior wherever possible. Native controls, real links, real headings, and real table elements carry accessibility, keyboard, and platform behavior that custom markup often breaks.

State and Variant Strategy

Every Sonner state should be reviewed before the component is marked ready. Start with the ordinary state, then check focus-visible, active, disabled, selected, loading, empty, and invalid states where they apply. A component that only has a polished default state will usually fail the first time it meets real product data.

For Sonner, the most important release checks are message is short, severity is accurate, and duration is appropriate. These checks should appear in examples, QA notes, and design review, not only in implementation code.

Variants should stay semantic. When a team asks for another color, size, or layout, ask what user meaning it communicates. If the answer is only visual preference, keep the extension local. If the answer is a repeated product state, promote it into the component contract with clear naming and examples.

Accessibility Requirements

• Keep messages concise.
• Avoid flooding live regions.
• Do not put critical form errors only in toast.
• Make undo actions reachable.

Accessibility for Sonner should be reviewed before visual polish is considered complete. The component must be reachable, understandable, operable, and recoverable. That means visible focus, meaningful names, correct roles, state communication, and a clear path out of mistakes.

The best accessibility test is not only automated. Use the keyboard. Increase browser zoom. Try a screen reader. Replace short demo labels with real production copy. These checks reveal whether Sonner survives the conditions users actually bring to the interface.

Responsive and Content Behavior

• Place toast where it does not cover primary controls.
• Keep width reasonable.
• Avoid stacking too many messages on mobile.

Avoid covering the primary task, keep messages short, and make recovery actions reachable with one hand. This is especially important for Tamil DS because translated or bilingual labels can expand quickly. Components should use flexible widths, stable spacing, and wrapping rules that keep content readable instead of compressing it into awkward fragments.

Responsive behavior is not just about screen size. It includes pointer type, zoom, reduced motion, high contrast, and slow network states. Sonner should remain understandable across all of those conditions.

Implementation Notes

Build Sonner as a reusable primitive first, then compose product-specific behavior around it. The base component should own structure, accessibility, state classes, and predictable spacing. Product screens can then add data, routing, analytics, and copy without reimplementing the core interaction.

Keep the rendered HTML close to the user's mental model. If the component performs an action, prefer a real button. If it navigates, prefer a real link. If it labels a form control, preserve the label relationship. These choices reduce bugs because the browser already understands common interaction patterns.

For maintainability, document the intended extension points. Teams should know which props are stable, which classes are safe to override, and which behaviors should stay centralized in Tamil Design System. That boundary keeps Sonner consistent while still allowing real products to move quickly.

Copy and Localization Notes

Sonner should work with direct English labels, longer Tamil labels, and bilingual product copy. Do not assume the demo text is the longest text the component will ever hold. Public-service products often need precise labels, district names, official terms, and explanatory helper text.

When writing copy for Sonner, prefer concrete nouns and verbs. Replace vague words like item, value, or action with the thing the user recognizes. If the component includes an icon, the text should still carry the meaning. The icon can reinforce the message, but it should not be the only explanation.

For design review, test Sonner with real examples from the product domain. If the component only works with short English sample text, it is not ready for Tamil DS documentation or production UI.

Design Review Questions

• Does Sonner solve the current user task more clearly than a simpler component?
• Are the visible labels specific enough to understand without surrounding explanation?
• Are disabled, selected, active, loading, empty, and error states designed before implementation starts?
• Does the component still work with long English labels, Tamil labels, zoomed text, and narrow mobile screens?
• Can a keyboard user reach, operate, and leave the component without losing context?

These questions help the team review Sonner as part of a real workflow. They also prevent the most common design-system failure: approving a polished default state while ignoring the states that users actually encounter.

Related Components and Alternatives

Choosing Sonner should be intentional. If the content is required to finish a task, the right answer may be another primitive, a plain content section, a dedicated page, or a clearer workflow. Component decisions are strongest when the team can explain both why this pattern was selected and why nearby alternatives were rejected.

Use the surrounding interface as the tie-breaker. A dense dashboard may need compact state treatment. A form may need clearer labels and validation. A navigation area may need route awareness. A feedback flow may need stronger recovery language. The same Sonner primitive can support these contexts, but the implementation should not hide the user's goal.

When a pattern starts accumulating exceptions, pause before adding more props. Repeated exceptions may reveal a new component, while one-off exceptions usually belong in the product layer.

Product Examples

• Saved successfully.
• Copied link.
• Upload failed.
• Undo archive.

Examples are useful because they make the component less abstract. When reviewing a Sonner implementation, ask which example it resembles. If the answer is unclear, the design may be mixing responsibilities from multiple patterns.

Common Mistakes

• Using toast for required instructions.
• Showing multiple toasts for one event.
• Auto-dismissing critical errors too quickly.
• Placing toast over mobile navigation.

Most Sonner mistakes come from treating the component as decoration rather than interaction design. The fix is usually to improve naming, reduce competing behaviors, expose state clearly, or move the task into a better component.

Production Checklist

• Message is short.
• Severity is accurate.
• Duration is appropriate.
• Action is reachable.
• Toast does not block controls.

Use this checklist as the release gate for Sonner. If one item fails, the issue is not cosmetic. It affects comprehension, accessibility, maintainability, or user trust.

During handoff, record the chosen variants, edge states, responsive behavior, and content assumptions. That note helps QA compare the implementation against the design instead of reviewing only the default state.