A static React application for comparing items side-by-side across shared attributes. Whether you're evaluating databases, programming languages, libraries, or any other category, Lineup helps you make informed decisions by presenting data in a clear, comparable format.

• Side-by-side comparisons - View multiple candidates simultaneously with their attributes aligned
• Grouped landing page - Comparisons organized into named groups, each tile carrying an optional decorative background graphic, with on-page sections explaining what Lineup is, how it's built, where the data comes from, and how to contribute
• Multiple comparison types - Support for databases, languages, libraries, frameworks, and more
• Rich value types - Display data as text, numbers, ratings, tags, icons, file sizes, durations, and more
• Collapsible attribute groups - Organize related attributes into logical sections
• Value ranking - Visual indicators for best/worst values based on configurable direction (higher or lower is better)
• Source citations - Every data point can include sources and comments, shown in tooltips on hover
• Responsive design - Works on desktop and mobile devices
• Self-contained HTML export - Export any single comparison to one portable .html file that renders and stays fully interactive (filter, sort, toggle candidates, dark mode) with no server, build step, or repo access — see Exporting a Comparison

• Node.js (v18 or later)
• pnpm package manager

# Clone the repository
git clone https://github.com/your-username/lineup.git
cd lineup

# Install dependencies
cd app
pnpm install

# Start development server
pnpm dev

The application will be available at http://localhost:5173.

cd app
pnpm build

The built files will be in app/dist/ and can be served by any static file server.

Export a single comparison type to one self-contained HTML file — everything (application JS, CSS, and the comparison's data) inlined, so the file opens directly in a browser with no server, build step, or repo access on the recipient's side:

# From the repo root
.scripts/export.sh databases > databases.html

# …or directly via the app package
pnpm --dir app export databases > databases.html

Open the resulting file in a browser: it server-side-renders on first paint, then hydrates into the fully interactive comparison view — filtering, sorting, candidate toggles, expandable rows, ranking/best-value highlighting, and prefers-color-scheme dark mode all work offline. The export reuses the app's own React components, so formatting and rankings always match the live app. The only possible network request is Devicon icon glyphs from a CDN; FontAwesome icons and all application logic and data are inlined. Passing an unknown comparison-type id fails with a clear error on stderr and a non-zero exit (no file is written to stdout). Interactive state is in-session only and is not persisted to the URL.

lineup/
├── app/                        # React application
│   ├── src/
│   │   ├── components/         # React components
│   │   ├── lib/                # Utilities
│   │   └── types/              # TypeScript types
│   └── package.json
├── data/                       # Comparison data
│   ├── index.json              # Comparison groups + list of available comparisons
│   └── <comparison-type>/      # One directory per comparison type
│       ├── index.json          # List of candidates
│       ├── attributes.json     # Attribute definitions
│       ├── tile.svg            # Optional decorative landing-page tile graphic
│       └── <candidate>.json    # Candidate data files
└── docs/                       # Documentation

Lineup ships with Claude Code skills under .claude/skills/ that handle each step of creating and maintaining comparison data. The skills are invoked as slash commands inside Claude Code.

Creating a new comparison type end-to-end:

1. Scope the type. /new-type <type> <optional seed> — iterate on the generated RESEARCH.md until purpose, scope, attribute groups, and candidates read well.
2. Scaffold schema and stubs. /scaffold-type <type> — generates attributes.json, registers the type in data/index.json, and creates empty stubs for each Initial Candidate.
3. Research each candidate. /gather-data <type> — auto-picks the next under-researched candidate, searches primary sources, and populates the values block. Repeat until every candidate is covered, or run /gather-data-cycle <type> [count\|all][@workers] to research the whole roster hands-off in parallel.

• Add a candidate you have in mind. /add-candidate <type> <name> <optional url/description/reason>, then /gather-data <type> <candidate-id> to research it.
• Discover candidates to consider. /discover-candidates <type> [hint] surfaces a vetted picklist from the web; picks are scaffolded automatically.
• Add a new attribute. /extend-comparison <type> <description>, then /gather-data <type> <candidate> <new-attribute-id> per existing candidate to fill the column.

The underlying layout is plain JSON and markdown — see Project Structure and the data schema in CLAUDE.md. The skills exist for convenience, not as a gate; direct edits work fine.

Lineup supports various value types for attributes:

{
  "name": "PostgreSQL",
  "description": "Advanced open-source relational database",
  "url": "https://postgresql.org",
  "values": {
    "license": {
      "value": "PostgreSQL License",
      "source": ["https://www.postgresql.org/about/licence/"]
    },
    "acid-compliant": {
      "value": true
    },
    "initial-release": {
      "value": 1996
    }
  }
}

• React - UI framework
• Vite - Build tool and dev server
• TypeScript - Type safety
• Tailwind CSS - Styling
• shadcn/ui - UI component library

pnpm dev             # Start development server
pnpm build           # Build for production
pnpm preview         # Preview production build
pnpm export <type>   # Export one comparison to a self-contained HTML file (stdout)
pnpm lint            # Run linter
pnpm format          # Format code

Contributions are welcome! Whether you want to:

• Add new comparison types with researched data
• Improve the UI/UX
• Fix bugs or add features
• Improve documentation

Please feel free to open an issue or submit a pull request.

MIT