4 ms·

Show HN: Knowing – an LLM tool built on concept hierarchies, not prompt-response

5 points by aponomy 2y ago

Hey HN! I've spent the past year full-time building Knowing, a tool for interacting with LLMs directly inside hierarchical structures instead of the usual prompt-response format. The idea started because I realized how much more intuitive it felt to build concept hierarchies continuously—no more endless copy-pasting or wondering how everything connects. The journey’s been a struggle. While I see huge potential in structuring AI interactions this way (writing books fast, planning projects, or organizing ideas), it’s been hard to pin down clear use cases in the market. I’m also working in near isolation on a small island in the Baltic Sea (born and raised here!), so any connections or feedback on direction would be a game-changer. I’d love feedback on a few things: - Does using structured hierarchies to interact with LLMs add value? - What use cases could resonate in the market? - Anyone here interested in helping me connect with others to bring this further? Try it here: https://knowing.app https://knowing.app (100 free credits with sign-up) Docs: https://docs.knowing.app https://docs.knowing.app Demo video: https://knowing.se https://knowing.se Thanks for any feedback or advice!

lunarcave 2y ago
Ok, this is cool. I use a few tools that does the same, but with CSVs. I often try to use "AI copilots for mind mapping", but this is good. Rather than trying to pin down the first potential use case, why don't you showcase a few use cases in the docs / home page that people can start as a template? - Insert your favorite book/author and explore peripheral works based on their other works / influences. - Learn X in 5 minutes - Take a concept, draw a mindmap and let people click around explore. If you can use some entity extraction to highlight the useful entities in a node, then you can build an experience where someone can generate the "next level" of a tree by clicking in a subject. I like the function part as well. Good luck! If you're ever in need of help, reach out. (E-mail in the profile)
- aponomy 2y ago
  Thanks! Appreciate the kind words and detailed feedback! Templates are definitely something I plan to add—it’s a great way to help users get started and explore use cases. I like the idea of exploring concepts gradually in a mindmap-style and making it possible to click to generate related ideas. The "next level" for expanding nodes etc. I wonder if adding these mindmap features are better at this point than other features I have in the pipeline, like automated workflows (to manipulate knowledge structures), attaching files to nodes, and even generating these files using different generative AI services and having information structures as argument. Lots of possibilities ahead. I’ll be sure to post an update here once I’ve made progress, especially on templates. Thanks again for the encouragement—it really means a lot!

5 points by aponomy 2y ago

lunarcave 2y ago

Ok, this is cool. I use a few tools that does the same, but with CSVs. I often try to use "AI copilots for mind mapping", but this is good. Rather than trying to pin down the first potential use case, why don't you showcase a few use cases in the docs / home page that people can start as a template? - Insert your favorite book/author and explore peripheral works based on their other works / influences. - Learn X in 5 minutes - Take a concept, draw a mindmap and let people click around explore. If you can use some entity extraction to highlight the useful entities in a node, then you can build an experience where someone can generate the "next level" of a tree by clicking in a subject. I like the function part as well. Good luck! If you're ever in need of help, reach out. (E-mail in the profile)

aponomy 2y ago
Thanks! Appreciate the kind words and detailed feedback! Templates are definitely something I plan to add—it’s a great way to help users get started and explore use cases. I like the idea of exploring concepts gradually in a mindmap-style and making it possible to click to generate related ideas. The "next level" for expanding nodes etc. I wonder if adding these mindmap features are better at this point than other features I have in the pipeline, like automated workflows (to manipulate knowledge structures), attaching files to nodes, and even generating these files using different generative AI services and having information structures as argument. Lots of possibilities ahead. I’ll be sure to post an update here once I’ve made progress, especially on templates. Thanks again for the encouragement—it really means a lot!