If an exact phrase appears in your data but doesn’t match in search, one possible cause is the position of that phrase within the attribute.
Exact phrase matching depends on word positions. Algolia only keeps the positions of the first 1,000 words of each attribute. If the phrase appears beyond that point, the engine may not be able to match it as an exact phrase, even when the text is present in the record.
This usually affects large text fields such as long descriptions, article bodies, or page content.
To reduce the impact of this limitation, the best approach is to index smaller records so the searchable content is split into shorter sections. For long documents or pages, this often means breaking the content into chunks instead of storing everything in one large attribute.
It can also help to review your searchableAttributes configuration so the most important content is prioritized in search. This won’t remove the 1,000-word position limit for phrase matching, but it can improve how relevant content is ranked and matched overall.
If your use case depends heavily on exact phrase queries, check whether the phrase is located deep inside a long attribute. If it is, restructuring the indexed content into smaller records is usually the most reliable solution.
Resources