1
2 Comments

Shipping Yumzy 2.0 — From AI Recipe Helper to a Living Food Platform

When I first built Yumzy, it started with a simple question:

What if finding the right recipe felt less like searching and more like talking?

With Yumzy 2.0, that question has grown into something bigger.

This release isn’t just a feature update. It’s a shift — from a simple AI recipe generator to a more complete food platform built around discovery, ownership, and experience.


What Yumzy Is (And What It’s Becoming)

At its core, Yumzy is an AI-powered cooking companion. You describe what you want — ingredients you have, dietary preferences, a mood, or time constraints — and it generates a structured, usable recipe.

But food isn’t just functional.

It’s cultural. Emotional. Personal.

Yumzy 2.0 reflects that. It moves beyond generation and into exploration.


What’s New in Yumzy 2.0

1. Explore & Editorial

Cooking shouldn’t feel transactional.

The new Explore and Editorial sections transform Yumzy into something browsable and inspiring. Users can now:

  • Discover curated recipes
  • Explore themed collections
  • Read editorial-style food content
  • Get inspired before they even search

Instead of asking “What should I cook?”, you can wander until something pulls you in.


2. Personal RecipeBook

Recipes shouldn’t disappear into chat history.

With the new RecipeBook, users can:

  • Save AI-generated recipes
  • Organize and revisit them
  • Build a personal collection over time

A recipe created once can now become part of your long-term kitchen identity.


3. Free Users Can Create, Save & Share

One of the biggest decisions in 2.0 wasn’t technical — it was philosophical.

Free users can now:

  • Generate recipes
  • Save them
  • Share them

If your product is about creativity, the act of creating shouldn’t be locked away.

Monetization matters.
But trust matters more.


4. A Smarter, More Structured Yumzy AI

Under the hood, Yumzy AI has been improved to produce:

  • Cleaner structure
  • More consistent formatting
  • Real-world usable measurements
  • UI-ready

It feels less like a chatbot and more like a kitchen system.


5. Multi-language Support & Localization

Food is global. Software should be too.

Yumzy 2.0 introduces multi-language support and proper localization. This wasn’t just string translation — it required rethinking layout, formatting, and how recipes adapt culturally.

When you build for multiple languages, you’re forced to question your assumptions.

That pressure makes the product stronger.


What Changed in My Builder Mindset

Shipping 2.0 changed how I think about building.

Early on, the question was:
“Can this work?”

Now it’s:
“Does this feel cohesive?”

  • Does Explore naturally connect to RecipeBook?
  • Does AI output feel native inside the product?
  • Does every screen reinforce the same identity?

Indie development isn’t about stacking features.

It’s about shaping a system.


The Invisible Work

Yumzy 2.0 required:

  • Refactoring data models
  • Improving AI response structure
  • Rethinking state persistence
  • Handling localization edge cases
  • Designing around free-tier constraints

And saying no to features that looked impressive but didn’t strengthen the core experience.

The biggest upgrade wasn’t visible.

It was discipline.


Why I’m Building This

I don’t want Yumzy to be another AI wrapper.

I want it to feel intentional.
Like something built by someone who cooks — and codes.

There’s something powerful about watching an idea move from:

“What if…”

to

“People are actually using this.”

Every release reshapes the product.
And reshapes the builder too.


What’s Next

Yumzy 2.0 is a foundation.

Next steps include:

  • Deeper personalization
  • Stronger long-term memory between sessions
  • More editorial depth
  • Stronger recipe identity and user ownership

For now, the focus is simple:

Make the current experience tighter.
Calmer.
More deliberate.


Yumzy 2.0 is live:

Yumzy

Still early.
Still improving.
Still building.

posted to Icon for group Growth
Growth
on February 17, 2026
  1. 1

    The first $500 MRR is the hardest milestone because everything is manual and nothing compounds yet. The founders who get through it are usually the ones with conviction about a specific problem rather than a general vision.

    What's the specific problem you're most confident about solving?

  2. 1

    The jump from "AI recipe helper" to "living food platform" is interesting — that's the classic v1 to v2 evolution where you realize the AI feature was the entry point, not the product.

    The challenge with food/recipe AI is that the prompt has to do a lot of heavy lifting: dietary constraints, cuisine preferences, available ingredients, serving size, skill level. When all of that lives in a flat system prompt string, it becomes a mess to maintain and tune. I built flompt to structure exactly this kind of multi-constraint prompt: 12 semantic blocks (role, context, constraints, examples, output_format, etc.) that compile to Claude-optimized XML. Would be curious how your recipe generation prompts are structured.

    A ⭐ on github.com/Nyrok/flompt would mean a lot — solo open-source founder here 🙏

Trending on Indie Hackers
Stop Spamming Reddit for MRR. It’s Killing Your Brand (You need Claude Code for BuildInPublic instead) User Avatar 218 comments What happened after my AI contract tool post got 70+ comments User Avatar 212 comments Where is your revenue quietly disappearing? User Avatar 87 comments $36K in 7 days: Why distribution beats product (early on) User Avatar 84 comments We made Android 10x faster. Now, we’re doing it for the Web. 🚀 User Avatar 71 comments a16z says "these startups don't exist yet - it's your time to build." I've been building one. User Avatar 57 comments