Wow. That is tremendous cool.

It could actually extract actual values (e.g. names, costs, dates), in addition to present ChatGPT-like semantic solutions (e.g. textual content abstract). Simply describe the entities with a easy format:

* description: a free textual content description of what you wish to extract.
* sort: string / float / integer / string.
* variable title: a descriptive variable title.
* (optionally available) legitimate values: restrict the output to a set of particular doable values.

Very spectacular, it labored nice on my dataset.

I just like the interactive demo on their web site ([https://www.textraction.ai/](https://www.textraction.ai/)) – it allowed me to strive my very own texts and entities inside minutes.

The service is accessible additionally as an API for any function by way of the RapidAPI platform: [https://rapidapi.com/textractionai/api/ai-textraction](https://rapidapi.com/textractionai/api/ai-textraction) (signal as much as RapidAPI and get your personal token)