recognizers-text

What it is and what it does

Recognizers-Text is a multilingual entity recognition library that identifies and resolves structured entities—numbers, ordinals, percentages, units (currency, temperature, dimensions, age), dates, times, email addresses, phone numbers, URLs, and more—from natural language text. It powers Microsoft's LUIS, Power Virtual Agents, and Bot Framework, and is also available as a standalone package.

The library targets multiple languages with varying levels of support: full support for English, Chinese, French, Spanish, Portuguese, German, Italian, Turkish, Hindi, and Dutch; partial support for Japanese, Korean, Arabic, and Swedish. It depends on emoji and multipledispatch for runtime operation. The Python package is currently in alpha status and has not been updated since November 2019, though the upstream repository remains active.

Use it for:

• Extract structured date and time expressions from user messages in chatbots or voice assistants.
• Parse currency amounts and unit measurements from product descriptions or user input.
• Identify and normalize phone numbers, email addresses, and URLs in text for data cleaning.
• Build multilingual NLP pipelines that need to recognize numbers, ordinals, and percentages across supported languages.
• Pre-process text for downstream ML models by extracting and normalizing temporal and numeric entities.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Recognizes and extracts entities like numbers, units, dates, and times from text in multiple languages, returning structured resolution data for each match.

Yes, with conditions. The package is permissively licensed and has low install friction, making it suitable for prototyping and integration into Microsoft ecosystem tools. However, it is alpha-stage and has not been updated since 2019; use it for production only if you can accept the maintenance risk and have validated that entity recognition quality meets your requirements for your target languages.

Install

recognizers-text on PyPI

pip install recognizers-text

uv add recognizers-text

poetry add recognizers-text

Installing recognizers-text

Before you install

Low friction install with only two runtime dependencies (emoji, multipledispatch). Package is in alpha status and has not been updated since 2019-11-12, though the upstream repository remains active.

License in practice

MIT license permits commercial and private use with minimal restrictions; you must retain the license notice in distributions.

Quickstart

pip install recognizers-text
from recognizers_text import recognize_number
results = recognize_number('I have two apples', 'en-us')

Package is alpha-stage and has not received updates since 2019; language support varies by entity type and culture.

Verify before relying

• Whether the alpha version is suitable for production use and what stability guarantees exist.
• Current state of partial language support (Japanese, Korean, Arabic, Swedish) and whether recognition quality is documented.
• Whether emoji and multipledispatch dependencies are required for all entity types or only specific recognizers.

Package facts

Evidence: recognizers_text-1.0.2a2-py3-none-any.whl

Keywords: nlp, nlp-entity-extraction, entity-extraction, parser-library

Tags

More Artificial Intelligence packages