Preview: Key phrase extraction model
[This topic is pre-release documentation and is subject to change.]
The key phrase extraction model identifies the main points in a text document. For example, given input text "The food was delicious and there were wonderful staff", the service returns the main talking points: "food" and "wonderful staff". This model can extract a list of key phrases from unstructured text documents.
Use in Power Automate
If you want to use this prebuilt model in Power Automate, you can find more information in Use key phrase extraction model in Power Automate.
Supported language and data format
- Documents can't exceed 5,120 characters.
- For information on language support, see Language and region support for the text analytics API.
If text is detected, the key phrase extraction model will output the following information:
- Results: A list of phrases from the document
- Phrase: Strings denoting the key talking points in the document text