S

Chat with PDFs, tool suggestions, text classifier

Chat with PDFs

You can now add PDFs to the chat, and ask questions about them.

  • Supports scanned documents. Advanced content extraction technology processes both scanned and digital PDFs (scanned files take a bit longer, ~5 sec/page).
  • Smart follow-up questions. Spark automatically suggests relevant questions based on the content of your PDF.
  • 100 pages per PDF. Only the first 100 pages of longer PDFs will be processed.
chat with PDF

Tool suggestions

Spark now suggests relevant chat tools alongside the standard open follow-up questions. Both suggestions are based on your previous question and Spark's answer. This makes Spark an even more effective jumping-off point and sparring partner for exploring topics:

tool suggestions

Text classifier workflow

The text classifier is a new autonomous workflow in Spark. It lets you categorize text collections, using your own custom classifications (= taxonomies).

For example, you can develop business-relevant filtering that aligns with your product lines, R&D priorities, or competitive landscape. This is something that standard classification systems like CPC (a patent classification system) or GICS (an industry classification system) can't provide.

The classifier always provides explanations for the classes it assigns to texts. And you can download the results.

text classifier workflow