Lets AI assistants look up UI component APIs, design tokens, examples, imports, and validation for building NVIDIA-style web interfaces.
Also known as NVIDIA/elements, @nvidia-elements/cli
- Stars
- 119
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 15, 2026
- License
- Apache-2.0
What is NVIDIA Elements MCP?
NVIDIA Elements lets AI assistants look up UI component APIs, design tokens, examples, imports, and validation for building NVIDIA-style web interfaces.
Summary generated from the project's README
About NVIDIA Elements MCP server
Building web interfaces for AI/ML workloads, robotics, or autonomous vehicles becomes more efficient when AI assistants can access specific design system data. This MCP server provides an agent-ready harness that exposes component APIs, design tokens, and import instructions directly to your terminal or AI client. It is designed to help developers implement NVIDIA-style interfaces across various environments.
Because the underlying web components are framework agnostic, you can use the information provided to build applications in React, Angular, Vue, Svelte, Lit, or plain HTML. The server handles the heavy lifting of looking up validation rules, code examples, and setup requirements. Developers can integrate this capability by adding the elements skill via the skills CLI to connect the design system directly to their AI-driven development workflows.
What you can do with it
- 1Retrieve component API detailsAsk your AI assistant to find the specific API properties for a UI component to ensure correct implementation.
- 2Implement design tokensRequest the exact design tokens needed to maintain visual consistency with the NVIDIA design system.
- 3Verify UI framework compatibilityCheck if a specific component can be used within your existing frontend stack like Vue or Svelte.
- 4Automate UI pattern validationUse the server to guide the authoring of best practices and common UI patterns during development.
Written from the project's README. Every use case is checked against a quote from it.
How to install NVIDIA Elements MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall NVIDIA Elements MCP in Claude Code
$ claude mcp add elements -- npx -y @nvidia-elements/cliInstall NVIDIA Elements MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"elements": {
"command": "npx",
"args": [
"-y",
"@nvidia-elements/cli"
]
}
}
}Install NVIDIA Elements MCP in VS Code
.vscode/mcp.json
{
"servers": {
"elements": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@nvidia-elements/cli"
]
}
}
}Install NVIDIA Elements MCP in VS Code CLI
$ code --add-mcp '{"name":"elements","type":"stdio","command":"npx","args":["-y","@nvidia-elements/cli"]}'Install NVIDIA Elements MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"elements": {
"command": "npx",
"args": [
"-y",
"@nvidia-elements/cli"
]
}
}
}Install NVIDIA Elements MCP in Codex
$ codex mcp add elements -- npx -y @nvidia-elements/cliInstall NVIDIA Elements MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.elements] command = "npx" args = ["-y", "@nvidia-elements/cli"]
Install NVIDIA Elements MCP in Gemini CLI
$ gemini mcp add -s user elements npx -- -y @nvidia-elements/cliInstall NVIDIA Elements MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"elements": {
"command": "npx",
"args": [
"-y",
"@nvidia-elements/cli"
]
}
}
}Automated checks
Ran Nov 12, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablePass
- Open source license (OSI)Pass
- Commit or release in the last yearPass
- Package name not close to a popular packagePass
- Remote endpoint uses HTTPSNot checked
- Destructive tools are labelledNot checked
- No hidden instructions to the modelPass
- No critical advisory in the package or its dependenciesNot checked
README
View on GitHubNVIDIA Elements
NVIDIA Design System and UI Agent Harness for AI/ML Factories, Robotics, and Autonomous Vehicles.
- Agent-ready tooling: CLI and MCP expose component APIs, tokens, examples, imports, validation, and setup to terminals and AI assistants.
- Framework agnostic: Web Components run in React, Angular, Vue, Svelte, Lit, plain HTML, server-rendered templates, and mixed stacks.
- Built for AI infrastructure: Operational UI for AI/ML workloads, autonomous vehicle tools, and robotics consoles.
- Stable API contracts: Skills and lint guide authoring best practices, common UI patterns, and automated static analysis.
NVIDIA Elements Skill
Install the Elements agent skill with the open skills CLI:
npx skills add https://github.com/nvidia/elements --skill elements
Requests and Contributions
- NVIDIA Elements Documentation
- Contribution Guidelines
- Feature request
- Bug report
- Official NVIDIA Elements DESIGN.md
Organization
The repository uses a top-level repository with individual project directories.
Project directories have their own package.json and commands. Run CI and development setup at the root repository level.
Examples of projects include:
/projects/starters- Suite of standardized starter apps for Elements and Patterns/projects/core- Elements library: curated UI maintained by the Elements team/projects/themes- Elements Theme library: provides a set of supported themes for Element based projects/projects/styles- Elements Styles library: provides a set of CSS utilities for layout and typography
Development
Setup
To set up repository dependencies and run the full build, run the following commands at the root of the repository:
The CI pipeline also builds Go starters. Install Go 1.26.x before running the full local CI pipeline.
# install dependencies https://mise.en.dev/getting-started.html
curl https://mise.run | sh
~/.local/bin/mise run setup
Troubleshooting
If you are coming from development from a different repository, you may need to refresh the repository toolchain with mise. Run mise run install at the root of the project to install the Node.js, pnpm, Vale, Go, Git LFS, and package versions from mise.toml and pnpm-lock.yaml.
If you actively switch between different repositories, run commands through mise exec -- or activate mise in your shell so it switches tools automatically.
Building
Both the top-level repository and each project have a set of standardized npm scripts. To build and test all projects, run mise exec -- pnpm run ci at the root of the repository.
Top-Level Repository
ci: run full build/lint/testci:all: entire CI process: build, lint, unit/lighthouse/visual testsci:reset: clear all caches/dependencies then reinstall dependencies
Individual Projects
Common project scripts include:
dev: run in watch modebuild: run project/library buildtest: run unit teststest:lighthouse: run lighthouse performance teststest:visual: run playwright visual regression teststest:axe: run axe tests for a11y
The available scripts vary by project. Check the project's package.json before running project-specific commands.
For details about the repository build and runtime flow, see the build system documentation.
Workflow
Before creating a branch or pull request, make a new issue or feature request first so the team can check alignment and avoid duplicate work.
Create a Branch
Use a descriptive branch name with the topic/ prefix. Example topic/bug-fix.
git checkout -b topic/bug-fix
After creating your branch, make your source code changes. After you complete them, run mise exec -- pnpm run ci in the root of the repo to run all the builds and tests. If all tests pass, you are ready to create a PR.
Commit Messages
The repo uses Semantic Release to manage package changes. Commit messages determine the release on merge. Commit Lint enforces commit formatting.
git commit -a -s -m "fix(core): disable multi-select"
| Types | Description |
|---|---|
fix |
bug fixes, performance fixes |
feat |
new features, components, APIs |
chore |
non production code modifications, build tooling, documentation |
| Scopes | Description |
|---|---|
ci |
CI and release automation |
cli |
/projects/cli |
code |
/projects/code |
core |
/projects/core |
create |
/projects/create |
deps |
dependency updates |
docs |
documentation and site content |
forms |
/projects/forms |
internals |
/projects/internals |
lint |
/projects/lint |
markdown |
/projects/markdown |
media |
/projects/media |
monaco |
/projects/monaco |
pages |
/projects/pages |
pi |
/projects/pi |
starters |
/projects/starters |
styles |
/projects/styles |
themes |
/projects/themes |
Keep commit names focused on your changes. The release process uses the commit message to determine the next release and generated changelog notes.
Opening a Pull Request
Once you have committed your changes to your branch locally, push them to the remote GitHub repository.
git push --set-upstream origin topic/bug-fix
Open a new Pull Request in GitHub. Request review from the team members and apply the appropriate labels in the GitHub UI, for example, type:fix and scope(core).
Amending Commit
If there are changes requested, make the requested changes locally and amend the commit.
git commit -a --amend --no-edit
This adds the changes to your existing commit. Then push the updated commit back to the remote branch for review.
git push --force origin topic/bug-fix
Rebasing Commit
If main changes before PR approval, rebase your local branch to include the latest changes from main.
git checkout main # Switch to main branch
git pull # Pull down any new changes
git checkout topic/bug-fix # Switch back to your topic branch
git rebase main # Rebase your branch onto the latest main
You may have to resolve any merge conflicts that arise from this process. Once complete, push the updated branch back to the remote repository for review.
New Project
When creating a new project, ex: ./projects/code, make sure to add the project to the pnpm-workspace.yaml located at the root directory.
Release
After approval, merge your Pull Request into main through the GitHub UI. GitHub Actions triggers a new release automatically. Semantic Release calculates the version from the commit type. The changelog also includes the commits from the PR.
NVIDIA Elements MCP questions
How do I add NVIDIA Elements MCP to Claude Code?
Run claude mcp add elements -- npx -y @nvidia-elements/cli in your terminal. Add --scope user to make it available in every project.
How do I add NVIDIA Elements MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What does the NVIDIA Elements MCP server expose to AI assistants?
It exposes component APIs, design tokens, examples, imports, validation, and setup to terminals and AI assistants.
Which frameworks does NVIDIA Elements support?
Elements are Web Components, so they work in React, Angular, Vue, Svelte, Lit, plain HTML, server-rendered templates, and mixed stacks.
How do I install the Elements agent skill?
Run the skills CLI command to add the skill from the NVIDIA elements repository.
What is NVIDIA Elements built for?
It provides operational UI for AI/ML workloads, autonomous vehicle tools, and robotics consoles.