♠️

buni 1-pager

What you’re building

buni is an AI-powered web app builder (that builds itself!)

Who it’s for

Initially (first few months):

  • buni speeds up workflow for a frontend dev. Does 80% of the work for you.
  • Best for people starting new projects or interested in modifying an existing template.

How it plays in the existing landscape

vs Cursor, Copilot

Short term: frontend app generation

  • buni is highly opinionated:
    • solely focused on web dev vs general-purpose programming
    • open by default
      • Enables introspection of buni core and outputs by other people and LLMs
    • better for de novo exploration instead of building on top of your existing codebase
  • buni is batteries-included:
    • hosting, version control, auth, database, APIs

vs Wix, Squarespace, Webflow

Medium term: integrate LLM capabilities

  • buni builds apps (reactive, data-driven), not just websites
  • buni will support technical LLM workflows (text, imagegen, agent architecture)

vs Twitter, Substack, Slack

Long term: widely distributed ecosystem/network of interoperable consumer & “LLM” apps

  • ecosystem & moat come from the new data format that underlies buni
  • buni probably integrates a native currency, similar to Manifold’s mana

notes

  • Matt: “this is where you share apps”, apps as a temporary, ephemeral thing seems pretty interesting
    • Many things that would have been text/pictures should be an app. Thinking of apps as cheap primitives.
    • [p] Similar to the tweet sent — generative AI enables more creation than things that used to be more permanent
  • Pippa: more concrete articulation of who it’s for
    • You’re using precious time to frontload your building, you’ll learn
    • Better to build it into v1, rather than doing more work unwinding
  • Matt: Depending on where you want to take it
    • Is there an initial thing that a lot of people want to make, the thing you want to double down on
    • Blank page problem blank textbox, what you solve?
      • Be opinionated
    • Templates: easy ways to understand
      • Send out survey to see what kind of development — within the first few weeks of
  • [a] Unclear if the early EF is the best cohort
    • [matt] Maybe first use case has to be “funner”. Ephemeral sharing thing
    • User group might not be people who have huge demands
      • Maybe more like “giphy” for apps — creativity, self expression
    • User personal of giphy for apps? idk, would choose a community
  • Matt: just ship something! No downside, just make it so that people can play with it, see what they do.
  • Pippa: Sooner you can ship the better
    • Hm, think Monday seems good
  • Matt: Thing that’s missing and adding would help: Step after you generate
    • If it’s going to be fun, shareability feels pretty key
    • Figure out some mechanism that makes it more seamless to share
      • [a] Yeah it’s embeddable — embedding in text or whatsapp
      • Slack would be interesting
      • Discord?
  • Mechanisms for virality, would be hard to go viral just as a website, easily enable people in a network
  • Who does visual design — internal Nick
    • Nick Williams
    • Send us what we’re doing on Monday
      • Something about shareability
        • Why Matt likes the analogy with Twitter or Substack. If you own the feed, you can own the experience.
          • Maybe not worth it