Download the PHP package yannkost/craft-asset-cleaner without Composer

Asset Cleaner for Craft CMS

Identify and clean up unused assets in Craft CMS 5.

Features

• View Usage Button - See where any asset is used directly from the asset edit page
• Utilities Page - Scan selected volumes for unused assets with bulk actions
• Staged Background Scanning - Scans run as chained queue jobs with dedicated setup, relations, content, and finalize stages
• Configurable Scan Storage - Choose between file-based and database-based scan storage
• File-Backed Scan State - Long-running scans can store progress and intermediate data under @storage/asset-cleaner/scans/<scanId> or a shared custom workspace path
• Database-Backed Scan State - Containerized and cloud-style installs can store scan state in the database when shared filesystem access is not guaranteed
• Configurable Draft & Revision Handling - Control whether draft-only and revision-only usage should count during scans
• Relational Fallback Option - Optionally treat any row in Craft’s relations table as usage for safer scans in projects with plugin-defined or unknown element types
• Large Library Friendly - Asset snapshots are chunked and content is processed in batches for much better performance on large datasets
• Per-Volume Results - Results grouped by volume with individual file counts and total sizes
• Export Options - Download CSV or ZIP of unused assets with smart filenames
• Folder Structure Option - Choose to preserve folder structure or flatten files in ZIP downloads
• Soft Delete (Trash) - Move assets to trash for safe removal with recovery option
• Permanent Delete - Permanently delete assets with double confirmation
• Progress Indicator - Visual progress bar during scanning, including stage-aware progress messages
• Memory-Efficient ZIP Creation - Handles large files without loading them fully into memory
• CLI Commands - Scan, export, and delete from the command line

Requirements

• Craft CMS 5.0.0 or later
• PHP 8.2 or later

Installation

Via Composer

Manual / local path repository

1. Copy the plugin to your plugins/ directory, for example to plugins/asset-cleaner

2. Add a path repository to your project's composer.json:

3. Require the package and install the plugin:

Usage

Control Panel

1. Navigate to Utilities → Asset Cleaner
2. Select the volumes you want to scan
3. Choose whether to Include drafts in this scan and/or Include revisions in this scan
4. Choose whether to Count all relational references as usage
5. Click Scan Now
6. Wait for the background scan to complete
7. Review the results grouped by volume, showing:
   • Number of unused assets per volume
   • Total file size per volume
8. Use bulk actions (global or per-volume):
   • Download CSV - Export a list of unused assets
   • Download ZIP - Download unused assets, optionally preserving folder structure
   • Put into Trash - Soft delete assets
   • Delete Permanently - Permanently remove assets with confirmation

The draft and revision toggles on the utility page override the plugin's default draft/revision handling settings for that individual scan. The relational fallback checkbox is scan-specific and is enabled by default for safer cleanup scans.

Scan stages

The utility page shows progress while the scan moves through these stages:

1. Preparing asset snapshot - Builds the stored snapshot of assets in scope
2. Scanning relations - Collects asset usage from Craft's relations table
3. Scanning content - Scans relevant rich text fields and globals in batches
4. Finalizing results - Merges usage data and writes the final unused asset results

ZIP Download Options

When downloading a ZIP file, you can choose:

• Flat - All files in the ZIP root
• Preserve folder structure - Files organized by volume-handle/folder/path/filename

Smart Filenames

Exported files include meaningful names with timestamps:

• CSV: unused-assets_volume-name_2024-01-15_14-30-00.csv
• ZIP: unused-assets_volume1__volume2_2024-01-15_14-30-00.zip

View Usage Button

On any asset edit page, click the View Usage button to see:

• Entries using the asset via Asset fields
• Entries referencing the asset in Redactor/CKEditor fields
• Global sets referencing the asset in supported rich text fields

CLI Commands

How It Works

The plugin identifies unused assets by performing several checks across your Craft CMS site:

1. Relations Table - Asset field references stored in Craft's relations table
2. Content Fields - Scans Redactor and CKEditor fields for embedded asset references
3. Global Sets - Checks supported global set content fields for asset references
4. Nested Entries - Resolves nested entries to top-level entries for usage reporting

For large scans, the plugin uses a staged pipeline:

1. Snapshot assets in scope into the active scan storage backend
2. Build usage sets once from relations and content instead of rescanning all content for every asset
3. Process entries in batches rather than loading the full content dataset into memory
4. Persist final unused results so completed scans can be restored and exported

The matching strategy prefers stronger signals first:

• data-asset-id references
• #asset:<id> references
• normalized URL/path matches
• unique filename fallback matches

The plugin can store transient scan state either on disk or in dedicated database tables, depending on your installation requirements.

Technical Details

Scan storage modes

Asset Cleaner supports two scan storage modes:

File-based storage

By default, scans are stored in a file-backed workspace under:

You can override the workspace base path with either:

• the ASSET_CLEANER_SCAN_PATH environment variable
• a scanWorkspacePath value in config/asset-cleaner.php
• the plugin settings in the control panel

Example plugin config:

Typical file-based scan artifacts include:

• meta.json
• progress.json
• asset chunk files
• relation/content usage ID files
• final unused asset results

File-based storage is a good default on single-server installs or setups where web and queue workers share the same filesystem path.

Database-based storage

Database-based storage keeps scan state, asset snapshots, used asset IDs, and final results in dedicated database tables instead of the filesystem.

This mode is recommended for:

• containerized environments
• cloud-style installs
• setups where web and queue workers do not share a reliable filesystem

You can switch between file-based and database-based storage in the plugin settings, and config file values can override those settings when needed.

Draft and revision handling

Asset Cleaner lets you control whether assets referenced only in drafts or only in revisions should count as used.

There are three levels of control:

• Config override - includeDraftsByDefault / includeRevisionsByDefault in config/asset-cleaner.php, or the ASSET_CLEANER_INCLUDE_DRAFTS / ASSET_CLEANER_INCLUDE_REVISIONS environment variables
• Plugin setting - sets the default behavior for new scans when no config override is present
• Utility page toggles - let you override those defaults for an individual scan

Example config override:

By default, draft-only and revision-only usage are excluded unless you enable them.

Relational fallback

Scans also provide a Count all relational references as usage checkbox.

When this option is enabled, any asset with a row in Craft’s relations table is treated as used. This is the safest mode for projects that rely on plugin-defined or unknown element types that may store asset relations outside normal entry content.

When this option is disabled, Asset Cleaner uses stricter relation resolution rules to identify more potentially unused assets, but that can undercount legitimate backend-only usage from third-party element types.

Scan performance tuning

The relation stage of a scan runs in resumable queue batches, and two settings control its footprint. Both are available in the plugin settings, and can be overridden via config or environment variables:

• Relation batch size (relationBatchSize in config/asset-cleaner.php, or the ASSET_CLEANER_RELATION_BATCH_SIZE environment variable, default 2000) - the maximum number of assets loaded for relation scanning per queue execution. Lower this (e.g. to 500) on sites with heavy or deeply nested relations if scan jobs time out.
• Relation time budget (relationTimeBudgetSeconds in config/asset-cleaner.php, or the ASSET_CLEANER_RELATION_TIME_BUDGET environment variable, default 120 seconds) - a wall-clock budget for the relation stage of a single queue execution. Once exceeded, the job stops and re-queues itself to continue, keeping each execution safely under the queue's time-to-reserve (TTR, 300 seconds by default). Keep this comfortably below your TTR.

Note that the effective relation batch size is max(assetChunkSize, relationBatchSize) - the scan never loads fewer assets per relation batch than its asset chunk size, so very low relationBatchSize values are clamped up to the chunk size.

Example config:

Bulk relation resolution

Relation sources are resolved in bulk with set-based queries, which reduces scan queries by orders of magnitude on relation-heavy sites. If you suspect it misbehaves on your install, you can fall back to the original per-source resolution path with:

• 'bulkRelationResolution' => false in config/asset-cleaner.php, or
• the ASSET_CLEANER_BULK_RELATION_RESOLUTION=false environment variable

You can verify on your own install that both paths reach identical verdicts with:

The command compares bulk and per-source verdicts over sampled assets in both relation-counting modes and reports any mismatch.

Memory-efficient ZIP creation

ZIP exports are built in a memory-friendly way:

• Files are streamed in chunks
• Temporary files are used instead of loading entire assets into memory
• Large files can be exported without exhausting PHP memory

Permissions

The Asset Cleaner utility requires the utility:asset-cleaner permission. Assign this permission to user groups that should have access to the utility.

License

The Craft License. See LICENSE.md for details.

Support

• GitHub Issues
• Source / Documentation

Changelog

1.5.1

Fixed database scan storage setup for fresh installations and upgrades. New installations now create all scan storage tables, and an upgrade migration repairs missing tables while preserving existing tables and scan data. Apply Craft's pending migrations after updating.

See CHANGELOG.md for the full release history.