Uber

Base

Public design system

Uber says Base defines the foundations of user interfaces across its products and services, and brings all Uber experiences together under a single, unified framework.

Shortcut points to it for

Multi-product scale
Extensions
AI-written design specs

Used for: Uber's ecosystem of products and services
Last checked: Verified 8 Oct 2026
Last updated: The site showed "Styleguide updated 5 days ago" when read.

Uber BaseOfficial source(opens in a new tab)

What Shortcut could read. The section overviews are public. Detail pages, including Design tokens, Color 2.0, What's new and About Base 2.0, ask for Uber's staff login, so Shortcut describes the structure and not the guidance inside.

What makes it interesting

Craft guidance

It separates a core from extensions. Maps, Rider and AI each extend Base instead of forking it.
Components are grouped by what they do for the user, such as feedback and status, not by widget type.
Haptics sits in the design language beside colour and type, which reflects a product used one-handed and on the move.
One spec has to serve seven implementation stacks, according to Uber's own engineering blog.

What designers can learn from it

How to let product teams extend a system without diluting it: a named extension library with its own playbook.
How to organise components by purpose, so people find them by the problem they have.
That a system this large still marks parts as coming soon. Publishing structure before content is normal.

Explore

System principles: Content principles, product inclusion, content heuristics. Design principles are marked coming soon.; Public
Design language: Border, colour, corner radius, design tokens, dimensions, elevation, layout grids, typography, haptics, icons, motion, illustrations, content.; Partly behind login
Components: Grouped as action, input and control, data display, feedback and status, surfaces, navigation, data and tables.; Partly behind login
Patterns: Feedback, inputting data, modality, states.; Partly behind login
Extension library: Maps, Rider, an AI design-guidelines hub, and a playbook for writing extensions.; Partly behind login
Tokens: A Design tokens page exists under Design language, at version 2.0.; Staff login

Accessibility

Uber's article on Base says a complete component spec includes a screen-reader section covering VoiceOver, TalkBack and ARIA, and describes its accessibility requirements as strict.

Official source

Uber Base
Uber Blog, How Uber Built an Agentic System to Automate Design Specs in Minutes(opens in a new tab)
Authority
Design-system convention
Published
11 Mar 2026
Last verified
Verified 8 Oct 2026

Responsive and device guidance

Base

Responsive guidance behind staff login

The pages that would cover this need a staff login, so Shortcut cannot say what the system publishes.

AI and the design system

How a design system becomes context for AI. Shortcut’s summary of these case studies.

Design system
Tokens
Components
Behaviour
Documentation
AI context
Generated or reviewed UI

Uber: an agent that writes component specs in Figma

The design-system lesson

Craft guidance

A design system is no longer only documentation for people. Uber's agent is useful because the system underneath is structured: real token names, real variants, a spec template. The AI contributes interpretation; the structure contributes accuracy. Without the second, the first has nothing to stand on.

What they say they did (6 points)

Uber's design systems team built uSpec, which connects an AI agent in Cursor to Figma through the open-source Figma Console MCP.
The agent reads the real component: structure, tokens, variables and styles. It then writes finished spec pages into the Figma file.
Each part of a spec has its own agent skill: anatomy, API, properties, colour annotation, structure and screen reader.
The screen-reader skill loads platform accessibility references first, so the agent selects from documented properties and does not guess.
Uber says a screen-reader spec covering iOS, Android and web generates in under two minutes, and that the pipeline runs locally so design data stays inside its network.
The specs serve seven implementation stacks.

Official source

Uber Base
Uber Blog, How Uber Built an Agentic System to Automate Design Specs in Minutes(opens in a new tab)
Authority
Design-system convention
Published
11 Mar 2026
Last verified
Verified 8 Oct 2026

What this means for generative UI

Explore

System principles

Content principles, product inclusion, content heuristics. Design principles are marked coming soon.

Public

Design language

Border, colour, corner radius, design tokens, dimensions, elevation, layout grids, typography, haptics, icons, motion, illustrations, content.

Partly behind login

Components

Grouped as action, input and control, data display, feedback and status, surfaces, navigation, data and tables.

Partly behind login

Patterns

Feedback, inputting data, modality, states.

Partly behind login

Extension library

Maps, Rider, an AI design-guidelines hub, and a playbook for writing extensions.

Partly behind login

Tokens

A Design tokens page exists under Design language, at version 2.0.

Staff login

Accessibility

Uber's article on Base says a complete component spec includes a screen-reader section covering VoiceOver, TalkBack and ARIA, and describes its accessibility requirements as strict.

AI and the design system

How a design system becomes context for AI. Shortcut’s summary of these case studies.

Design system
Tokens
Components
Behaviour
Documentation
AI context
Generated or reviewed UI

Uber: an agent that writes component specs in Figma

The design-system lesson

Craft guidance

What they say they did (6 points)

Uber's design systems team built uSpec, which connects an AI agent in Cursor to Figma through the open-source Figma Console MCP.
The agent reads the real component: structure, tokens, variables and styles. It then writes finished spec pages into the Figma file.
Each part of a spec has its own agent skill: anatomy, API, properties, colour annotation, structure and screen reader.
The screen-reader skill loads platform accessibility references first, so the agent selects from documented properties and does not guess.
Uber says a screen-reader spec covering iOS, Android and web generates in under two minutes, and that the pipeline runs locally so design data stays inside its network.
The specs serve seven implementation stacks.

Official source

Uber Base
Uber Blog, How Uber Built an Agentic System to Automate Design Specs in Minutes(opens in a new tab)
Authority
Design-system convention
Published
11 Mar 2026
Last verified
Verified 8 Oct 2026

What this means for generative UI