Full Copies and Recovery

A full copy, called a snapshot in the API, is the complete current row list for one public resource.

For example:

• Full copy of mushafs:1 returns its metadata, page mappings, and positioned words. Font files and images are not included.
• Full copy of translations:19 returns all current translation rows for translation resource 19.
• Full copy of word_by_word_translations:85 returns all current word translation rows for word-by-word translation resource content 85.
• Full copy of word_by_word_transliterations:60 returns all current word transliteration rows for production word-by-word transliteration resource 60.
• Full copy of tafsirs:151 returns all current tafsir rows for tafsir resource 151.
• Full copy of recitations:10 returns ayah audio file rows and compatible chapter audio file rows for legacy ayah recitation 10.
• Full copy of chapter_recitations:159 returns chapter audio file rows for Audio::Recitation.id 159.
• Full copy of articles:123 returns visible article localization rows for article 123.

When You Need a Full Copy

RESOURCE_CREATE and RESOURCE_INVALIDATE include snapshot_url. Fetch it before storing the new sync token.

Full Copy Response

The response has metadata plus records.

{
  "resource_group": "translations",
  "resource_id": 19,
  "resource_content_id": 19,
  "schema_version": 1,
  "sync_sequence": 98234,
  "records": [
    {
      "id": 85108,
      "verse_key": "26:153",
      "text": "They said: Thou art but one of the bewitched;"
    }
  ]
}

Client rule: records is the full current content for that one resource. Delete your old local rows for the resource, then insert these rows.

Word-by-Word Transliteration Records

The word_by_word_transliterations group is separate from the translation-only word_by_word_translations group. Its records use record_type: "word_transliteration" for incremental row mutations and are ordered by word_id, then id in snapshots.

For production resource 60, a record has exactly these fields:

{
  "id": 1,
  "resource_content_id": 60,
  "resource_id": 60,
  "word_id": 60,
  "language_id": 38,
  "language_name": "english",
  "text": "bis'mi",
  "updated_at": "2026-08-19T02:43:00Z"
}

resource_id == resource_content_id == ResourceContent.id identifies the owning resource, while word_id identifies the Word owner. Replace the complete local resource when bootstrap or a RESOURCE_INVALIDATE mutation supplies a snapshot_url; apply ROW_CREATE, ROW_UPDATE, and ROW_DELETE by the mutation's resource_group, resource_id, record_type, and record_key.

Chapter Recitation Records

Chapter-recitation snapshots use the same envelope as other snapshots, but their records contain only chapter audio files:

{
  "resource_group": "chapter_recitations",
  "resource_id": 159,
  "resource_content_id": 415,
  "schema_version": 1,
  "sync_sequence": 98234,
  "records": [
    {
      "record_type": "chapter_audio_file",
      "id": 2001,
      "audio_recitation_id": 159,
      "chapter_id": 1,
      "audio_url": "https://example.com/audio/001.mp3"
    }
  ]
}

The resource_id is the Audio::Recitation.id; it is not the legacy Recitation.id used by the recitations group. The legacy group may also receive chapter-file mutations for backward compatibility, but its snapshot is not the independently addressable chapter-recitation resource.

Unavailable Resources

Public Content Sync is not an editorial audit log. It tells public clients what they need to display current public content.

Server change	Client action
Resource becomes hidden, rejected, unapproved, or deleted	RESOURCE_DELETE: remove or hide the local resource.
Hidden resource changes while unavailable	No row changes are replayed to public clients.
Resource becomes public again	RESOURCE_CREATE: fetch the full copy and replace local rows.
Resource needs a complete refresh	RESOURCE_INVALIDATE: fetch the full copy and replace local rows.

Error Recovery

Error code	Meaning	Client recovery
resync_required	Token or cursor cannot be used.	Discard local token and bootstrap this resource filter again.
token_filter_mismatch	The token belongs to a different resource filter.	Use the matching filter or bootstrap the new filter.
cursor_filter_mismatch	The cursor belongs to a different resource filter.	Continue with the original next_page_url, or restart the sync.
cursor_per_page_mismatch	The cursor was used with a different per_page.	Use the returned next_page_url path without changing per_page.
snapshot_not_found	The resource full copy is not public or no longer exists.	Treat the resource as unavailable locally, then continue syncing.

Re-Bootstrap Flow

For endpoint details, see Get content resource snapshot.