We’re always evolving Grooper to make it not only more powerful, but also easier to use. Our newest updates are a perfect example of that vision: they’re designed to cut setup time, boost accuracy, and let AI handle the complexity.
In this article, we’re spotlighting two features you might not have heard about yet, but will want to use immediately.
Let’s talk document separation and classification, now powered by Large Language Models (LLMs).
Auto-Separate: Now with Built-In Classification
In the last minor build of 2025, we introduced Auto-Separate, a groundbreaking method for document separation that uses an LLM to follow plain-language instructions.
For example:
“Create a new document every time the invoice number changes.”
That’s all it takes. Simple natural-language instructions to distinguish a new document. No inserted separator pages, no patch codes, no barcodes. Just AI doing the heaving lifting, as it should be.
But now, we’ve made it even more powerful: Auto-Separate can also classify documents as it separates them.
Here’s how it works:
- Add one or more Document Types to the “Document Types” property.
- The LLM will automatically choose the correct Document Type for each separated document.
- The LLM uses not just the name of the Document Type, but also its description field to make more informed choices.
In simpler setups, the name alone may be enough. But for more nuanced workflows, adding clear, distinct Descriptions gives the model better context and better accuracy.
No Document Types listed? No problem, Auto-Separate will still do its separation work, just without classification.
Classification That Reads Between the Lines
Want LLM-powered classification without separation? You’ve got it.
We’ve added a brand-new LLM Classifier method to the Classification Provider. This lets Grooper classify standalone documents by prompting an LLM to pick the best Document Type from your Content Model.
Just like with Auto-Separate, the model considers:
- The name of each Document Type.
- The Description of each Document Type, if provided
It’s simple. It’s powerful. And in many cases, it outperforms traditional methods, especially when document meaning matters more than matching exact words.
Why This Matters
We’re not just adding AI for the sake of AI. We’re building features that make document automation:
- Easier to configure
- Faster to deploy
- Smarter out of the box
With Grooper’s LLM-based Auto-Separate and LLM Classifier, you can now streamline complex workflows without writing a single rule. These features give you the freedom to let the model do the work while you focus on results.
These updates are live and ready. Try them out, and let us know what you discover.
Your feedback shapes how we keep pushing the boundaries of intelligent document automation.
