deep-dive

What Kind of Database Does Hallmark Use? The Answer Revealed

August 3, 2026 Nutlope/hallmark ↗

Hallmark uses no database at all. It is a purely static front-end tool that generates self-contained HTML, CSS, and JavaScript files without any server-side persistence layer.

If you're analyzing the Nutlope/hallmark repository to understand its data architecture, you'll find something unexpected: Hallmark operates entirely without a database. This design is intentional—the project exists to build standalone, shareable UI artifacts that require no backend infrastructure to function.

Hallmark's Static-Only Architecture

Unlike typical web applications that rely on PostgreSQL, MongoDB, or SQLite for data persistence, Hallmark takes a radically minimalist approach. The codebase contains no ORM configurations, no connection pools, and no query builders—because there's nothing to persist beyond the files themselves.

Examining package.json reveals the complete dependency picture:

{
  "scripts": {
    "serve": "npx serve site"
  },
  "devDependencies": {
    "serve": "^14.2.0"
  }
}

The only dependency is serve, a lightweight development HTTP server. There are no database-related packages such as:

  • sqlite or sqlite3
  • mongoose or mongodb
  • prisma
  • sequelize
  • pg (PostgreSQL driver)
  • redis

According to the hallmark source code, the project is built with zero runtime dependencies that would enable data persistence.

How Hallmark Stores "Data"

Hallmark doesn't store user data—it generates static artifacts. The site/index.html file demonstrates this architecture:

<!doctype html>
<html lang="en" data-theme="hum">
  <head>
    <meta charset="utf-8" />
    <title>Hallmark – A design skill that refuses to look AI‑generated</title>
    <link rel="stylesheet" href="css/tokens.css" />
    <link rel="stylesheet" href="css/base.css" />
  </head>
  <body>
    <header class="banner">…</header>
    <main class="page">…</main>
    <footer class="slot slot--foot"></footer>
    
    <!-- External analytics only -->
    <script defer data-domain="hallmark.dev" src="https://plausible.io/js/script.js"></script>
  </body>
</html>

All content exists as hardcoded HTML and CSS. There's no fetch() call to an API, no localStorage usage for structured data, and no dynamic rendering that would require a backend database.

The One External Service: Plausible Analytics

The only third-party service referenced in site/index.html (lines 93-99) is Plausible Analytics—a privacy-focused, hosted analytics platform. This is worth clarifying because:

  • Plausible is not a database used by Hallmark
  • It performs anonymous traffic collection on the hosted version
  • It requires no configuration, schema, or integration code in the repository
  • Self-hosted Hallmark builds function completely without it

What This Means for Deployment

Hallmark's database-free design enables unique deployment patterns:

Deployment Target Requirement
GitHub Pages Works immediately—just static files
Netlify/Vercel Zero configuration needed
Any CDN Direct file serving, no origin compute
Local filesystem Open index.html in a browser

The skills/hallmark/SKILL.md documentation emphasizes this constraint: Hallmark generates self-contained UI artifacts that must function without any supporting infrastructure.

When This Architecture Matters

Understanding what database Hallmark uses (none) helps clarify when to adopt this tool versus alternatives:

  • Use Hallmark when you need shareable, portable design artifacts with guaranteed longevity
  • Avoid Hallmark when you need user-specific state, dynamic content, or collaborative editing

The tradeoff is explicit: Hallmark sacrifices persistence for reliability and zero-maintenance deployment.

Summary

  • Hallmark has no database—it's a pure static site generator producing HTML, CSS, and JavaScript files
  • package.json contains zero database dependencies; only the serve dev server is present
  • site/index.html confirms the architecture: no server-side code, no API calls, no data persistence
  • The only external reference is Plausible Analytics (third-party, optional, not a database)
  • All "data" exists as static markup in generated files

Frequently Asked Questions

Does Hallmark use SQLite for local storage?

No. Hallmark does not use SQLite or any embedded database. The repository contains no sqlite3 imports, no .db files, and no SQL queries. Everything exists as plain text in HTML and CSS files.

Could Hallmark work with a database if I added one?

Technically yes, but this would violate the project's core design. The SKILL.md documentation frames Hallmark as a generator of self-contained artifacts. Adding a database would require forking the project and implementing an entirely new architecture.

How does Hallmark handle user preferences without a database?

Hallmark appears to use no persistent user state based on the source code. Any theming (seen in data-theme="hum") is hardcoded in the generated HTML. There's no evidence of localStorage usage for settings persistence.

Is Plausible Analytics a database that Hallmark uses?

No. Plausible is a hosted third-party service for anonymous traffic analytics. It receives data from visitors but does not function as Hallmark's data store. The project works identically with Plausible removed.

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"
Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →