TypeUI DESIGN.md Extractor
- Overview
The TypeUI DESIGN.md Extractor is a Chrome Extension designed to capture the visual and structural signals of a live website and convert them into reusable design documentation. Its core purpose is to provide a reliable blueprint in the form of DESIGN.md or SKILL.md that can be consumed by design-to-development workflows and AI-assisted tooling such as Google Stitch, Claude Code, Codex, and other agents. By translating on-page tokens like typography, color palettes, spacing scales, radii, shadows, motion characteristics, and interaction hints into a formalized format, the extension helps teams maintain design consistency across sites and products.
This tool aligns with the open-source TypeUI DESIGN.md specification. The output follows the conventions defined by the TypeUI DESIGN.md format, ensuring compatibility with existing tooling and ecosystems that rely on standardized design-system narratives. By adhering to this format, the extracted information becomes interoperable, enabling teams to bootstrap design systems and documentation without manual transcription.
A central visual reference is provided to illustrate the kind of output and context users may expect. The screenshot included in the input depicts the interface and the resulting markdown artifacts in action, offering a tangible sense of how signals map to the generated files. The image serves as a visual cue for the user journey—from active tab extraction to the production of DESIGN.md or SKILL.md.
The extractor emphasizes a structured approach to design documentation. Rather than producing free-form notes, it aims to organize insights into clearly defined sections and tokens, supporting consistent interpretation by designers, developers, and AI agents. The resulting files are intended to be “agent-ready,” meaning they can be ingested by automation pipelines, code generators, or design-review processes with minimal manual tweaking.
In essence, the extension functions as an accelerator for design-system blueprints. It empowers teams to rapidly capture and codify a site’s design language, then export those signals into formats that accelerate implementation, governance, and collaboration across disciplines.
The output documents emphasize a disciplined workflow, balancing design intent with practical guidance for implementation. The resulting DESIGN.md or SKILL.md not only describe what the design looks like but also how it should be applied, tested, and extended within real-world projects. This dual focus—descriptive clarity and actionable guidance—helps bridge the gap between design intent and code-level realization.
Note on accessibility and reliability: the design tokens and rules captured by the extractor are intended to support accessible implementation. The produced content can be used to drive WCAG-adherent interfaces, consistent semantic usage, and predictable interactions that teams can verify against established accessibility criteria.
The extension is positioned as a practical companion for teams adopting a design-system-centric workflow. By converting live site characteristics into formalized markdown artifacts, teams gain a reusable reference that can be version-controlled, shared, and evolved over time as design requirements shift or expand.
The following sections detail how to get started, what actions are available, what the generated file structure looks like, how to develop or test the extension locally, and the licensing terms that govern its use.
[Image: Screenshot 2026-04-16 at 21 27 09]
- The embedded screenshot illustrates the user interface and the result of an extraction, providing a concrete depiction of how typography, color, spacing, and tokens translate into the DESIGN.md or SKILL.md content. This image is sourced from the project’s assets and serves as a reference point for users exploring the extension’s capabilities.
- Getting started
Load the extension into Chrome to begin creating design-system blueprints from live sites. The process is straightforward and mirrors standard extension installation steps, enabling developers and designers to adopt the tool quickly in their existing workflows.
Step 1: Open chrome://extensions in the Chrome browser. This is the central hub where installed extensions are listed and managed, and where you can add new or updated extensions.
Step 2: Enable Developer mode. This toggle enables advanced features required for loading unpacked extensions, including the ability to load a local folder project directly into Chrome for testing and development.
Step 3: Click Load unpacked. This action opens a file picker to select a directory that contains the extension’s source code and manifest. Choosing the correct folder ensures the extension is loaded with all necessary resources and configuration.
Step 4: Select this project folder. By choosing the folder that contains the extension’s source, you activate the development version of the tool. This setup enables you to test extraction, generation, and other actions against live pages, validating signals and the resulting markdown outputs.
After completing these steps, the extension becomes usable within the Chrome environment. Users can navigate to any site, prompt the extension to auto-extract signals, and then generate DESIGN.md or SKILL.md documents based on the captured data.
A practical note: once loaded, you can use the extension to refresh the current page state. This re-runs the extraction against the active tab, ensuring that the generated outputs reflect the most up-to-date visual tokens and interaction cues observed on the site.
The getting-started flow emphasizes a minimal, deterministic onboarding that aligns with common Chrome extension practices, reducing friction for teams that routinely prototype or document design systems.
- Curated design skills
The extension points users toward curated design systems and resources that complement the extraction workflow. For teams exploring design tokens, themes, and system-wide guidelines, the curated collection at TypeUI Design Skills provides a curated gateway to validated patterns, components, and design-language concepts.
The integration with curated design skills is intended to enrich the design-system blueprint by connecting the extracted signals to established design principles, tokens, and best practices. This alignment helps ensure that the generated DESIGN.md or SKILL.md not only mirror the site’s surface appearance but also reflect broader, reusable patterns that can be scaled across products.
Access to curated resources is beneficial for teams seeking to standardize terminology, token naming conventions, and recommended implementation patterns. It supports a coherent vocabulary that developers and designers can share as part of the organization’s design-system governance.
The curated link provided in the input points to a dedicated design-skills resource that users can consult in parallel with extraction, enabling more robust and well-structured documentation. This synergy between live-site signals and curated design guidance strengthens the quality and consistency of the final markdown artifacts.
- Available actions
Auto-extract: The extension can read styles from the active tab, capturing a comprehensive array of design signals. These signals include typography attributes (font families, sizes, line heights, letter spacing), color values (primary, secondary, neutral palettes), spacing metrics (margins, paddings, grid gaps), radii (border radii for components), shadows (surface depth and blur), and motion characteristics (transitions and timing).
Generate DESIGN.md: From the extracted signals, the tool synthesizes a design-system documentation file in markdown format. This file encapsulates the design language, tokens, and implementation guidance in a structured narrative suitable for design reviews, developer handoffs, and tooling pipelines.
Generate SKILL.md: The extractor can produce an agent-ready skill markdown, designed for AI-assisted workflows and automation. This document emphasizes actionable guidance for agents or automation scripts to reproduce or apply the observed design patterns in new contexts.
Refresh: Trigger a re-extraction for the current page state. This is useful when you have made changes to the page, loaded a new section, or want to ensure the latest signals are captured before generating the final outputs.
Download: Save the generated output as DESIGN.md or SKILL.md. This creates a local file that can be version-controlled, shared with teammates, or integrated into automated pipelines for documentation and implementation workflows.
Explain (?): View an explanation of how the file was generated, drawing on the TypeUI reference. This feature provides transparency into the extraction process, showing the signals that were detected and how they map to the produced markdown sections.
The action set is curated to cover the end-to-end lifecycle of converting live-site signals into structured documentation, from signal extraction to artifact generation and distribution.
- Generated file structure
The generated markdown adheres to a formal structure designed to capture both the design intent and the practical guidelines needed to implement that intent. Each section serves a specific purpose to ensure comprehensive coverage of the design system for the site being analyzed.
Mission: This section defines the overarching objective of the design system for the extracted site. It frames the problem the design language aims to solve and contextualizes the site within a broader product strategy.
Brand: In this section, the extension captures product and brand context, the site’s URL, target audience, and the product surface. It anchors the design decisions in brand essence and user expectations, ensuring a coherent voice and visual identity across implementations.
Style Foundations: This part lists inferred visual tokens and foundations, including typography scales, color ramps, spacing systems, radii, elevation tokens, and motion parameters. It documents the building blocks that underlie the site’s visual language and are reusable across components.
Accessibility: The extension emphasizes WCAG 2.2 AA requirements and interaction constraints, outlining accessibility considerations tied to the extracted design signals. This section guides developers on ensuring keyboard navigability, adequate color contrasts, focus management, and other accessibility practices aligned with the observed design.
Writing Tone: Here, guidance on tone and voice helps ensure that content and UI copy align with the design language. It sets expectations for how text should feel and read within the implemented interfaces, enabling consistent user experiences.
Rules: Do: A set of recommended implementation practices that should be followed to faithfully reproduce the design language. These are positive guidelines focusing on correct token usage, component behavior, and alignment with the design system.
Rules: Don't: This section enumerates anti-patterns and prohibited behaviors to avoid drift from the intended design language. It helps prevent shortcuts that could degrade accessibility, consistency, or visual fidelity.
Guideline Authoring Workflow: Defines the ordered steps for creating and maintaining guidelines. It outlines a repeatable process for authoring, reviewing, and updating documentation as the design or site evolves.
Required Output Structure: Enforces consistency across outputs by specifying mandatory sections and their order. This ensures that every DESIGN.md or SKILL.md produced by the extension follows a predictable schema, making automation easier and reviews more efficient.
Component Rule Expectations: Defines required interaction and state details for components. This section clarifies how components should respond to user input, what states they should present (hover, focus, active, disabled), and how they should communicate their status to assistive technologies.
Quality Gates: Adds testable quality and consistency checks. These gates help verify token presence, naming conventions, accessibility constraints, and alignment with the defined design-system rules, providing a measurable standard for validation before publication.
The generated file structure is designed to be human-readable while also machine-parseable, enabling both designers and automation to leverage the documentation effectively.
- Local development
The project supports local development and testing to ensure that changes behave as expected before publishing or distributing the extension. Running the test suite locally helps maintain quality and reliability as the project evolves.
To run tests locally, execute the following command in your terminal:
node tests/run-tests.mjs This command uses the project’s test runner to execute the test suite, validate functionality, and catch regressions early in the development cycle.
Local development typically involves iterative cycles of editing extension code, loading the unpacked extension into Chrome (as described in the getting started section), performing manual extractions on sample sites, and running automated tests to confirm expected behavior. The emphasis is on reproducibility and determinism so that the outputs remain consistent across runs.
Documentation and example pages are often used during development to illustrate the expected structure of DESIGN.md and SKILL.md artifacts. By following the defined file structure and signals, developers can verify that generated outputs align with the TypeUI DESIGN.md format and meet the quality gates described earlier.
The local development workflow is designed to be approachable for both frontend engineers and UX-focused contributors, ensuring that a broad range of team members can contribute to extending or improving the extractor’s capabilities.
- License
This project is open-source under the MIT License. The MIT License is widely used for permissive software distribution, allowing reuse, modification, and redistribution with minimal restrictions. It promotes collaboration and adoption while preserving attribution to the original authors.
Readers and contributors should review the license text to understand any obligations related to attribution, liability, and warranty disclaimers. The permissive nature of MIT makes it a good fit for developers who want to integrate the extractor into broader toolchains, internal workflows, or commercial products without imposing heavy licensing constraints.
As with many open-source projects, the MIT License encourages community involvement, contributions, and transparency. Users who benefit from the extractor can propose enhancements, report issues, and contribute fixes or new features, helping the project mature over time.
The license choice aligns with the open-source ethos presented by the TypeUI DESIGN.md ecosystem and supports adoption by teams of varying sizes and domains who seek to automate and standardize their design documentation processes.
- Practical usage patterns and considerations
Workflow integration: The extractor is designed to fit into common design-to-code pipelines. By exporting DESIGN.md or SKILL.md, teams can wire outputs into version control, documentation dashboards, or AI-assisted tooling that consumes design-system blueprints. This approach helps unify design intent and implementation guidance across projects.
Token fidelity and interpretation: The accuracy of extracted tokens depends on the live site signals and the extension’s interpretation rules. The goal is to capture faithful representations of typography, color, spacing, radii, and motion. When signals are ambiguous or context-dependent, the generated documentation may include clarifying notes or guidelines to resolve interpretation differences during implementation.
Accessibility alignment: Observing WCAG 2.2 AA requirements in the generated Accessibility section helps teams ensure that color contrast, focus states, and keyboard navigation are considered from the outset. The documentation can serve as a checklist for developers to verify accessibility as features are implemented.
Design-system governance: The Required Output Structure and Quality Gates sections are intentionally structured to support governance. By enforcing consistent sections, token naming conventions, and validation checks, organizations can maintain a cohesive design-language narrative across multiple sites and teams.
Extensibility: Because the output is markdown-based and aligned with an open-design format, it is naturally adaptable. Teams can augment the DESIGN.md or SKILL.md with organization-specific tokens, components, or guidelines while preserving the core structure that makes automation and cross-project reuse straightforward.
Screenshot and visual references: Including visual references, such as the embedded screenshot in the input, helps maintain a bridge between the live site and the generated documentation. Visual cues can support reviewers who are assessing whether the extracted design language matches the intended appearance and interaction behavior.
- Summary of benefits
Accelerated design-to-development handoff: The extractor converts live-site visuals into structured design-system markdown, providing a clear starting point for developers and AI agents to reproduce the design language with fidelity.
Consistency and scalability: By codifying tokens, rules, and guidelines, teams can scale their design language across pages, components, and products while preserving consistency.
AI-assisted workflows: The agent-ready SKILL.md format enables AI systems to leverage the extracted signals for code generation, component proposals, or automated validation against the design system.
Open-format compatibility: The reliance on the open-source TypeUI DESIGN.md format ensures compatibility with a broader ecosystem of tools and resources, facilitating collaboration and interoperability.
Transparency and traceability: The Explain feature offers visibility into how a file was generated, promoting trust and enabling teams to audit and adjust the extraction process as needed.
- Closing notes
The TypeUI DESIGN.md Extractor is presented as a practical facility for teams seeking to anchor their websites in a formal design-language blueprint. By combining live-site extraction with a standardized markdown structure, the extension supports robust documentation, reproducible handoffs, and collaborative workflows that bridge design, development, and automation.
The project embraces an open-source philosophy via the MIT license, inviting contributors, researchers, and practitioners to participate in its evolution. Whether you are a designer mapping tokens, a front-end engineer implementing a system, or an AI enthusiast building automation around design languages, this tool aims to streamline processes, reduce manual translation work, and foster consistent design outcomes.
The included screenshot serves as a tangible reminder of the extension’s workflow, illustrating how on-page signals transform into structured outputs. As you explore the extension, you will encounter the balance between fidelity to the source site and the clarity of the resulting documentation—the ultimate aim being a reliable, reusable blueprint that supports efficient design-system governance and real-world implementation.
Enjoying this project?
Discover more amazing open-source projects on TechLogHub. We curate the best developer tools and projects.
Repository:https://github.com/bergside/design-md-chrome
GitHub - bergside/design-md-chrome: TypeUI DESIGN.md Extractor
The TypeUI DESIGN.md Extractor is a Chrome Extension designed to capture the visual and structural signals of a live website and convert them into reusable desi...
github - bergside/design-md-chrome