README

PHP-native build-time full-text search indexer with a dependency-free JavaScript query runtime. No Node, no server-side query API, no Elasticsearch — index at build time, query in the browser.

Why Trawl?

Trawl fills a specific gap: PHP shops that want client-side full-text search without adding Node to their toolchain.

	Trawl	Pagefind	Hosted (Algolia, etc.)
Indexer runtime	PHP (Composer)	Node/npx	SaaS
Search runtime	Browser-only JS	Browser-only JS	Server API
Node dependency	None	Required for indexing	None
Composer install	`composer require --dev`	N/A	N/A
Chunked index	Yes (CRC32 shards)	Yes	N/A
On-premise	Fully self-contained	Fully self-contained	Third-party

If you already have Node in your build pipeline, Pagefind is excellent and more mature. If you don't, or you want your entire dev dependency tree in Composer, Trawl is your option.

Quickstart

composer require --dev commonform/trawl
trawl init
trawl build --source=dist/ --output=dist/trawl/

Then in your HTML:

<script type="module">
import { Trawl } from '/trawl.js';

const trawl = await Trawl.init({ baseUrl: '/trawl/' });
const results = await trawl.search('reverse proxy');
// [{ url, title, score, excerpt: '...<mark>reverse proxy</mark>...' }]
</script>

Copy runtime/trawl.js from the package into your public directory, or use it directly from vendor/commonform/trawl/runtime/trawl.js.

CLI Commands

trawl build   Build the search index
trawl verify  Validate manifest + shard + doc consistency
trawl watch   Watch for file changes and rebuild incrementally
trawl init    Scaffold a trawl.json configuration file

Build

trawl build --source=dist/ --output=dist/trawl/
trawl build --adapter=markdown --source=content/ --output=dist/trawl/
trawl build --incremental                          # skip unchanged docs
trawl build --config=trawl.json                    # use config file

Verify

trawl verify --index=dist/trawl/

Configuration (trawl.json)

{
    "source": "dist/",
    "output": "dist/trawl/",
    "adapter": "filesystem",
    "include": ["**/*.html"],
    "exclude": ["**/404.html", "drafts/**"],
    "selectors": {
        "title": "title",
        "heading": "h1, h2, h3",
        "body": "main",
        "ignore": "[data-trawl-ignore], nav, footer"
    },
    "field_weights": { "title": 10, "heading": 5, "body": 1 },
    "facets": ["tags", "category"],
    "language": "en",
    "stemmer": "porter2",
    "stopwords": "default",
    "incremental": true,
    "cache_file": ".trawl-cache.json"
}

Adapters

Adapter	Input	Description
`filesystem`	`.html` files	Walks a directory of rendered HTML, extracts content via CSS selectors
`markdown`	`.md` / `.mdx` files	Parses YAML frontmatter + Markdown body via league/commonmark
`array`	`iterable<Document>`	Programmatic input from CMS integrations

JS Runtime API

import { Trawl } from './trawl.js';

const trawl = await Trawl.init({ baseUrl: '/trawl/' });

const results = await trawl.search('query', {
    limit: 10,
    filters: { tags: 'docs' },
    mode: 'and'               // 'and' (default) or 'or'
});

// results: [{ url, title, score, excerpt, meta }]

Requirements

Indexer: PHP 8.3+, Composer
Runtime: Any browser with native fetch, ES modules, and URL

License

MPL-2.0

commonform / trawl

Package info

Security

Statistics