Skip to main content

​
Overview

The Orama Database generator creates a searchable database using the Orama search engine. It indexes API documentation with hierarchical titles, descriptions, and site sections for fast client-side search.

​
Metadata

​
name
string
default:"orama-db"
Generator identifier
​
version
string
default:"1.0.0"
Generator version
​
dependsOn
string
default:"metadata"
Requires metadata generator output as input

​
Configuration

​
output
string
Output directory path for the generated database file
​
minify
boolean
default:"false"
Enable JSON minification of the database file

​
Output Format

The generator produces:
  • Database file: orama-db.json
  • Format: Orama database format with indexed documents
  • Schema:
    • title: Hierarchical title (e.g., “fs > readFile”)
    • description: First paragraph of entry content
    • href: Link to entry (e.g., fs.html#read-file)
    • siteSection: Top-level API module name

​
Usage

doc-kit -t orama-db
With minification: 
doc-kit -t orama-db --config.orama-db.minify=true

​
Dependency Chain

ast → metadata → orama-db
The Orama database generator depends on:
  1. ast - Parses Markdown to AST
  2. metadata - Extracts API metadata
  3. orama-db - Indexes into searchable database

​
Implementation Details

The generator:
  • Creates an Orama database instance with predefined schema
  • Groups nodes by API module
  • Builds hierarchical titles showing the path to each entry
  • Extracts first paragraph as description
  • Inserts all documents in batch
  • Saves serialized database to JSON file

​
Schema Definition

The Orama database uses this schema (from constants.mjs): 
{
  title: 'string',
  description: 'string',
  href: 'string',
  siteSection: 'string'
}

​
Hierarchical Titles

The buildHierarchicalTitle function creates breadcrumb-style titles:
  • Shows the path from module to specific entry
  • Example: "File System > fs.readFile > Synchronous API"
  • Helps users understand context in search results

​
Description Extraction

The generator:
  • Finds the first paragraph node in entry content
  • Converts to plain text using transformNodeToString
  • Strips markdown formatting for clean descriptions
Source: src/generators/orama-db/

Build docs developers (and LLMs) love

Get started for freeTalk to us