For AI agents: this page is also available as Markdown at /docs/phone-numbers.md. Every docs page is: add .md to its URL. llms.txt lists them all.

Phone numbers

The same number is written (415) 555-2671, +1 415 555 2671 and 415.555.2671, and people search for it by its last four digits. The phone analyzer reads a phone number in any common format and indexes it in its international form (+14155552671), its digits with and without the country code, and its trailing digits down to four. So all three forms match each other, and a search for 2671 finds the number. Extensions are ignored, and digits in other scripts, such as Arabic-Indic or fullwidth, are read as digits.

{"fields": {"phone": {"type": "text", "analyzer": "phone"}}}

Use the phone_search analyzer instead when only a whole number should match: it indexes the international form and the national number, without the trailing-digit forms.

#Other regions

Numbers written without a country code are read as US numbers. For another region, use the tokenizer with default_region:

{
  "type": "text",
  "analyzer": {"tokenizer": {"name": "phone", "params": {"default_region": "GB"}}}
}

#Notes

• A value that is not a valid phone number is indexed as its digits.
• The phone_digits filter keeps a leading + and the digits of each token and can follow any tokenizer.
• The OpenSearch phone and phone-search analyzers and tokenizers map to these, and phone-region sets default_region. See Through the compatible APIs.

View this page as Markdown