> ## Documentation Index
> Fetch the complete documentation index at: https://www.meilisearch.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Prefix search

> Prefix search is a core part of Meilisearch's design and allows users to receive results even when their query only contains a single letter.

In Meilisearch, **you can perform a search with only a single letter as your query**. This is because we follow the philosophy of **prefix search**.

Prefix search is when document sorting starts by comparing the search query against the beginning of each word in your dataset. All documents with words that match the query term are added to the [bucket sort](https://en.wikipedia.org/wiki/Bucket_sort), before the [ranking rules](/capabilities/full_text_search/relevancy/ranking_rules) are applied sequentially.

In other words, prefix search means that it's not necessary to type a word in its entirety to find documents containing that word—you can just type the first one or two letters.

Prefix search is only performed on the last word in a search query—prior words must be typed out fully to get accurate results.

Searching by prefix (rather than using complete words) has a significant impact on search time. The shorter the query term, the more possible matches in the dataset.

### Example

Given a set of words in a dataset:

`film` `cinema` `movies` `show` `harry` `potter` `shine` `musical`

query: `s`:
response:

* `show`
* `shine`

but not

* `movies`
* `musical`

query: `sho`:
response:

* `show`

Meilisearch also handles typos while performing the prefix search. You can [read more about the typo rules on the dedicated page](/capabilities/full_text_search/relevancy/typo_tolerance_settings).

### Disabling prefix search

You can disable prefix search entirely using the [`prefixSearch` index setting](/reference/api/settings/get-prefixsearch). Set it to `disabled` to turn off prefix search for an index. The default value is `indexingTime`, which enables prefix search.

<Tip>
  We also [apply splitting and concatenating on search queries](/resources/internals/concat).
</Tip>