@libremax_off Fond rend chaque recette computable par l’IA, ingrédient par ingrédient.
Meal plans instantanés, listes de courses intelligentes et paniers e-commerce deviennent automatiques.
https://t.co/0M1kOo6CKp
One cool thing about the @fond_kitchen ingredient engine:
When you import a recipe, it can automatically attach relevant tags based on what’s actually inside it.
Not just the title or description.
The ingredients.
The hardest part of building an AI cooking app is not the chat interface.
It is ingredients.
One recipe says “1/2 onion.” another says "1 chopped onion".
A meal plan needs quantity, category, pantry state, duplicates, substitutions, and grocery list merging.
That gap looks small.
It is basically the whole product.
I’m building a mascot system for Fond, my cooking app.
The goal isn’t just “make a cute character.”
It’s to create a mascot that can idle, react, celebrate, cook, wait, jump between states, and still feel like the same little character every time.
Here’s the workflow I’m using:
1. Character base
I start in @recraftai to explore the character style, poses, and personality.
Once the direction feels right, I create one neutral “idle” image of the mascot, centered, same scale, no background.
This becomes the source of truth.
2. Animation loop
Then I use @fal with:
fal-ai/kling-video/v3/pro/image-to-video
The key feature is first frame / last frame.
I use the same PNG as both the first and last frame, so every animation starts and ends in the exact same idle pose.
That means animations can loop cleanly, or transition from one state to another without weird jumps.
3. Chroma-key background
In the prompt, I specify:
“First and last frame must match exactly: neutral idle pose, centered, same scale, same camera, on a perfectly flat solid #00FF00 chroma-key green background.”
Why green?
Because transparent AI video usually leaves little dirty edges around the character. A clean green background gives me more control later.
4. Cleanup in @capcutapp
If the animation looks good, I open it in CapCut:
- Remove background
- Use chroma key
- Pick the green background
- Adjust intensity until the cutout looks right
- Create a compound clip
- Go to Adjust > HSL
- Select green
- Drop saturation to 0
And that’s it.
Now I have a reusable animated mascot clip that can work inside the app.
Still experimenting, but this first-frame / last-frame trick is the part that makes the whole system feel consistent.
Most AI cooking apps start with the wrong problem.
They try to generate more recipes.
But most people already have too many recipes.
The real problem is turning messy food intent into action:
what can I cook?
what do I need?
what step am I on?
what can I substitute?
That is the part I care about.
Love this. @VoriHQ is exactly the kind of infrastructure we'd want to plug into down the line.
We're building the AI ingredients layer at @fond_kitchen, turning any input into a formatted, ready-to-use recipe + grocery list (solving the "1.7 avocado" scaling problem).
Congrats on the raise 🚀
Most meal planning apps treat grocery lists like math homework.
Add a recipe and you get:
“2.5 avocados”
Cool. Let me just ask the cashier for half an avocado.
@fond_kitchen is built around a simple idea:
recipes are math, groceries are real life.
You need 3 avocados.
I’m building Fond because recipe apps still treat recipes like text.
But cooking software needs to understand ingredients:
- “lemon zest”
- “5cl lemon juice”
- “lemon slices”
That should become one lemon on your grocery list.
Fond turns messy recipes into structured ingredient data, then uses that to make planning, shopping, scaling, and AI actually work.
Pre-launch. Building in public from here.
https://t.co/TRN5xbnKjx @fond_kitchen
I post about my pizza dough calculator on X (@fond_kitchen) → 30 views, 0 likes.
I post one photo of the actual pizza on Reddit → 20k views, 250+ upvotes, 10 comments.
As a non-American, I completely underestimated Reddit's reach. Even a new account can explode instantly in the right sub.
Recipe: https://t.co/ByCPzWEewH
There are 200+ recipe apps out there.
We're building another one — and here's why it's different.
Most recipe apps work great if you cook Western-style meals, read English, and shop at a standard supermarket. But real cooking isn't like that.
It's your grandmother's mole from Oaxaca. Your neighbor's kimchi jjigae. That late-night TikTok pasta you saved. A tagine from a Moroccan food blog written entirely in French.
These recipes don't fit neatly into existing apps. We think they should.
That's why we built Fond.
Import a recipe from anywhere — a URL, a photo, a handwritten family card, even a foreign-language blog. The AI figures it out regardless of language or format.
Plan your week in minutes. One tap for a shopping list, no duplicates.
The recipe app space is crowded. But it's crowded with apps built for the same people.
We're building for everyone who cooks.
Early access is opening soon. If you're tired of apps that don't fit how you cook, join the waitlist -> https://t.co/zEh2r9to9H
Recipe writers need first-class tools. That's why @fond_kitchen is packed with an advanced rich editor.
The trick users don't know? It uses @cooklangorg under the hood.
Soon, non-technical people will use Cooklang without even knowing it.