Show HN: Turn Multiple Documents into Comparative Tables with GPT
tactic.fyi
Managing and comparing data across numerous documents is time-consuming. That's why we created Tactic Generate, an app that lets you transform scattered information from documents into a structured, tabular output.
One great use case in scientific research is performing a literature review where you pull in lots of papers to compare their methods, experimental results, and conclusions. These docs might all be written and structured differently, making the comparative process very manual. With Tactic Generate, literature reviews will be much faster and easier, potentially helping scientists spot patterns across many more papers than previously possible. Personally, I have an academic background as a battery materials researcher, and I wish I’d had a tool like this to help me extract and compare synthesis methods, chemical formulas, and electronic properties observed across a wide range of scientific literature.
This is just the tip of the iceberg. I think there are also lots of other applications in business domains we can apply this to.
This app lets you upload lots of papers and documents, ask questions about them, and get answers back in the form of a table, where we group the answers by document or topic. This allows you to use an LLM to ask the same question about lots of documents at the same time and get the answer for each document in a table, along with total summaries.
I think there is a lot of potential here in moving away from a pure, chat-like experience, and giving the user more control over how the LLM is used to extract information from a collection of underlying documents and what the display and output looks like.
The demo video on the landing page is targeted toward business applications, but you can swap out the business topics for scientific ones and it becomes very powerful for those subjects.
This is a super early alpha version of this product, but we wanted to get it out there and collect your feedback. Please let us know what you think.