A web app with traditional Asian recipes and an AI assistant to simplify cooking
Project gallery
Updated Jun 3, 2026
Trust score
79/90
Umami Guru is a web app for those who love Asian cuisine but hesitate to cook at home due to complex recipes and rare ingredients. The platform offers ready-made recipes with AI features: finding substitutes for ingredients available in local supermarkets, scaling portions, calculating the cost of the shopping basket, and a step-by-step assistant that guides you through each stage of cooking. The app helps overcome the barrier of strictly following a recipe by adapting dishes to what you have on hand.
How did Umami Guru start?
I have a collection of Asian recipes, and every time I want to cook something, I'm stopped by the need to strictly follow the ingredient list. If I don't have the right fish or vegetable, I have to find a substitute, but I also need to consider the specifics of cooking. Umami Guru solves this problem by finding alternatives in local stores and adapting the recipe to the new ingredient on the fly.
How is Umami Guru different from similar projects?
The closest alternative is Magic Kitchen, but our platform already contains ready-made recipes from which the user chooses what to cook. In the future, we might introduce a service for selecting dishes from the database based on available ingredients, but for now the focus is on ready-made recipes with an AI assistant.
What exactly did the neural networks do, and what remained for you as the author?
The neural networks wrote all the code for the AI features. I handled the design, the idea, and the overall architecture of the app.
How long did the current version take, and what was the longest part?
Three months with breaks for other projects. The longest part was designing the ingredient substitution logic and integrating the AI models.
What is currently the biggest obstacle to the next step?
I need to manually populate the platform with my recipes. I have a configured local handler for parsing text and images, but first I need to move hosting from Firebase to my own and replace the current AI models, as they have become too expensive.
Assessed Jun 7, 2026
A focused Asian recipe platform with integrated AI helpers, but similar services exist.
Similarity
Overlaps with EatWhat (Asian recipes + AI) and Magic Kitchen (AI from ingredients) by offering AI-assisted Asian cooking, though features differ.
Difference
Umami Guru provides ready curated recipes with AI for substitutions, scaling, and cost calculation, while alternatives focus on ingredient-to-recipe or general Asian recipe databases.
Closest matches
EatWhat — Ask AI What to Cook | 1,100+ Asian Recipes
EatWhat — Ask AI What to Cook | 1,100+ Asian Recipes The end of “what should I cook?” # What should I cook tonight? No more staring into the fridge. Tell EatWhat a dish, an ingredient, or a craving — and get a real recipe for tonight, sized to your table, with step-by-step timers. Free · Pro optional 6 languages Halal & veggie Calories & macros Free on iOS — your answer’s seconds away. No credit card. Eat what? Chicken wings, 30 min 🍗 Coke Chicken Wings — a Hong Kong classic. 385 kcal,
Magic Kitchen - AI Chinese Cuisine Master
Magic Kitchen - AI Chinese Cuisine Master # Magic Kitchen AI AI Chinese Cuisine Master - Transform your ingredients into delicious dishes. Simply input your available ingredients, and AI will recommend the most suitable Chinese recipes for you, making every meal a delightful surprise! 🧙♂️ 🔮🍳 Ready to experience AI magic? Transform ingredients into delicious dishes ✨ Cast Magic ✨ ## 🌟 Product Features Smart matching, precise recommendations, making cooking simple and fun 🤖 ### AI
Asian Chef-Free Asian cuisine recipes and resources.
# Asian Chef - Asian cuisine recipes and resources. [...] Asian Chef is an AI-powered virtual assistant specializing in Asian cuisine recipes. Designed to cater to culinary enthusiasts, home cooks, and anyone interested in exploring the diverse culinary traditions of Asia, Asian Chef offers a wide range of features and resources to assist users in their cooking endeavors. Whether you're a beginner looking to learn the basics of Asian cooking or an experienced cook seeking authentic recipes and c
Asian Recipes Cookbook - App Store
Asian Recipes Cookbook App - App Store # Asian Recipes Cookbook Smart Recipes & Tips Free · In‑App Purchases · Designed for iPad. Not verified for macOS. Share iPhone, iPad Organize your cooking with step-by-step recipes, and helpful kitchen tips. Save time, cook better, and enjoy delicious dishes every day with zero stress. Asian Recipes Cookbook – AI Cooking Assistant Welcome to Asian Recipes Cookbook, your all-in-one companion for cooking authentic, flavorful Asian dishes at home! AI-P
Tushar9802/bawarchi-recipe-generator · Hugging Face
# Bawarchi Recipe Generation Model [...] Role-aware fine-tuned Llama 3.2 3B for cuisine-context recipe generation [...] Bawarchi is a recipe generation model that understands ingredient roles and adapts recipes based on cuisine context. Fine-tuned from Llama 3.2 3B using LoRA with a novel curriculum learning approach over 335K recipes, plus role-aware extension training on 1K curated examples. [...] - Cuisine-context switching: Same ingredients → different dishes by cuisine (e.g., chicken + tort
The idea addresses a real problem — the barrier to entry for Asian cuisine. It includes AI features (ingredient substitution, portion recalculation, cost estimation) that differentiate it from simple recipe collections. However, the content is sparse, there is no marketing strategy, and competition from EatWhat and...
Strong sides
Clear niche: Asian cuisine with an emphasis on ingredient accessibility — a sought-after topic for Russian-speaking audiences.
AI features (ingredient substitution, portion recalculation, cost estimation) solve specific user pain points and set the project apart from basic recipe lists.
The author has personal experience and motivation — the project grew from a personal recipe collection, adding authenticity.
Weak sides
Very few recipes on the site (visually fewer than 10), which gives users no reason to return.
No marketing strategy or acquisition channels — the project relies solely on word-of-mouth.
The interface and design look minimalistic, lacking a clear focus on user experience (no filters, search, categories).
What to improve next
Focus on content: at least 30–50 recipes with high-quality photos and step-by-step instructions.
Develop a simple marketing plan: target Russian-speaking Asian cuisine enthusiasts, collaborate with food bloggers.
Improve UX: add filters for difficulty, cooking time, dish type; make the mobile version more user-friendly.
Growth ideas
Integrate affiliate links to products from local supermarkets — monetization and usefulness at once.
Create a content plan: publish 1–2 recipes per week with video tutorials on social media (Instagram, Pinterest) to drive traffic.
Add a "recipe of the day" or seasonal collections — increases engagement and return visits.
Card-to-product alignment
The project card and live site largely match: the described AI features are present, albeit in minimal form.
The interview confirms the author's personal story but lacks concrete plans for content or marketing.
Uniqueness is confirmed (score 14), but competitors (EatWhat, Magic Kitchen) already have more content and audience.
Need to manually populate the platform with recipes. There is a local handler for parsing text and images, but first I need to move hosting from Firebase to my own and replace the current AI models, as they have become too expensive.
Need help from a designer and marketer to develop the platform and attract an audience.
Unlock the temporary creator workspace before sending a report. This keeps report mutations tied to an accountable user until full auth is ready.
Open creator workspaceA quick jump to the creator's other published work on storevi.be.
Web application for renting construction machinery with AI assistant and CRM for small businesses in Austria
Built by lokki
Multilingual community for publishing and promoting projects created with vibe‑coding
Built by lokki