Forum Academy Marketplace Showcase Pricing Features

🗣 New Plugin - NLP Toolbox - Easily add Natural Language Processing to your App

Hi all!

We’re happy to announce that we’ve just released a new plugin to add Natural Language Processing (NLP) capabilities to your app. With it you can do:

  1. Sentiment Analysis - Determine if an input text is positive, negative, or neutral
  2. Summarization - Paraphrase a long text into a shorter one.
  3. Named Entity Recognition - Extract names of People, Places, and Organizations from text.
  4. Question Answering - Automatically answer questions about input text.

Try it out and let us know what you think! (links below) We’d love to hear your feedback. And if there’s an NLP feature you’d like added to the plugin, reach out to us; we’d be happy to chat about including it.

Thanks!
Korey @ Launchable AI

You can try out the plugin here: http://nlp-toolbox-demo-app.bubbleapps.io/version-test

You can see the editor view here: Nlp-toolbox-demo-app | Bubble Editor

Finally, the plugin page is here: NLP Toolbox Plugin | Bubble

5 Likes

Hey @launchable tried your plugin, and looks very promising, I might use this for my app soon.
really like the sentiment analysis.

1 Like

Thanks! Hope it’s helpful!

Hello there, can you give me information on the underlying engine that you’re using? Is it the Stanford open source project?

Hi @lucas.ar,

It’s not the Stanford project, no. The endpoints and processing are all built on custom code and infrastructure. We’re using a variety of models for the tasks, adopted from the NLP literature.

Hope this helps!

Would greatly appreciate if you could shed some light on it but understand if it is protected IP or something

What languages are supported?

Hi @lucas.ar

Right now only English is supported. We’ve been looking into on-the-fly translation to support other languages though. Are you thinking about a specific use case for another language, or for multiple languages?

Regarding the underlying models, we’re using adaptations of a few popular architectures, including:

And for pre-processing we’re using some task-specific tokenizers, mostly based on BERT.

Is that more helpful?

Cheers

2 Likes

Wow! This looks so Cool! Going to have a think about how a can use this.

@timgarrett111 - Great! Let me know if I can help!

I would really love to try it if it can support Arabic as it will be a great added value to my regulations search engine

Hi @ali.roshdy,

I’m going to look into supporting multiple languages. A few people have asked about this now, so it seems like there’s a need. I’ll let you know how things progress.

Cheers