Search

MinimalDoc includes client-side full-text search with fuzzy matching.

Usage

Keyboard Shortcut

Press Cmd+K (macOS) or Ctrl+K (Windows/Linux) to open search.

Click

Click the search input in the sidebar.

Navigation

Key	Action
↑ / ↓	Navigate results
Enter	Go to selected result
Escape	Close search

Search Index

A search-index.json file is generated during build:

[
  {
    "title": "Installation",
    "description": "How to install MinimalDoc",
    "url": "/getting-started/installation.html",
    "tags": ["getting-started", "installation"],
    "content": "First 500 characters of content..."
  }
]

Indexed Fields

Field	Weight	Source
Title	Highest	Frontmatter title
Description	High	Frontmatter description
Tags	Medium	Frontmatter tags
Content	Lower	Page body (first 500 chars)

Configuration

Search is enabled by default:

# config.yaml
enable_search: true

Disable search:

enable_search: false

Fuzzy Matching

Search tolerates typos and partial matches:

Query	Matches
instal	“Installation”
confg	“Configuration”
api ref	“API Reference”

Index Size

Typical index sizes:

Pages	Index Size
10	~5 KB
50	~25 KB
100	~50 KB
500	~250 KB

Optimization Tips

Good Titles

# Good - Specific, searchable
title: Configure OAuth 2.0 Authentication

# Avoid - Generic, hard to find
title: Configuration

Descriptive Tags

tags:
  - authentication
  - oauth
  - security
  - api

Front-Load Content

Put important keywords in the first 500 characters, as that’s what gets indexed.

Unique Descriptions

# Good - Unique, descriptive
description: Set up OAuth 2.0 authentication with Google, GitHub, or custom providers

# Avoid - Generic
description: Authentication documentation

Excluding Pages

Hidden pages are excluded from search:

---
title: Internal Notes
hidden: true
---

Search UI

The search interface includes:

• Search input with placeholder
• Real-time results as you type
• Result highlighting
• Keyboard navigation indicators
• Result count
• “No results” state

Technical Details

Algorithm

1. Tokenize query into words
2. For each word, find fuzzy matches in index
3. Score matches by field weight and match quality
4. Sort by combined score
5. Return top N results

Performance

• Index loaded once, cached in memory
• Search executes in <10ms for most sites
• No server round-trips