3
4 Comments

PH launched today! Threadfisher, an tool that replies in visuals instead of text.

I think the assumption that you need to know your chart type before you ask a question is backwards, and I wanted the visual-first default to exist before every AI tool bolts on a "chart mode" as an afterthought. Threadfisher sits between your question and the answer: you ask anything, even confused or half-formed, and a local rendering layer figures out the right visualization and builds it before replying. Every visual is interactive, click to explore, annotate, refine through follow-up, and export as HTML, PNG, or embed.

Built solo end to end, including the visualization system and the 3D rendering.

Live on Product Hunt today: https://www.producthunt.com/products/treadfisher?utm_source=other&utm_medium=social

posted toAvatar for product Threadfisher
Threadfisher
  1. 1
    That behavior change after a few uses is probably one of the more interesting signals here. The best AI interfaces may be the ones where users eventually stop thinking about how to structure the request and just describe the business question they’re trying to answer. I’d watch whether that shift also increases repeat usage, not just the complexity of the questions.
  2. 1
    The visual-first default is the interesting thesis, not the 3D rendering. Curious whether early users are actually asking questions differently when they expect a visual answer, or whether they still prompt like a normal AI tool and just appreciate the visualization afterward.
    1. 1
      Honestly still figuring that out. Early read: most people still prompt like they're talking to a normal AI tool at first, old habits. But the ones who stick around start asking messier, more exploratory questions ("why does this look off," "what am I missing here") once they realize they don't have to pre-specify a chart type. That shift seems to happen after 2-3 uses, not immediately. Did you get a chance to try it? Curious what you asked and what came back.
      1. 1
        That makes sense. The driver/courier focus seems like a useful way to build the density you need without trying to solve coverage everywhere at once.