Food changes feelings. (🤯)
You’ve seen a grumpy person eat, and suddenly the edges soften. You’ve also seen people drink a coffee and lock in; or calm down after a cup of tea.Â
We all try to influence our states with what we eat and drink, just not very intentionally.
I became “the friend who cooks” almost by accident. I used to host my friends, cook, watch people unwind, and post twice a year to a tiny Instagram page called Potsandfeelings. No master plan. just food and how it made people feel. That page quietly became my lab for design, writing, restaurant reviews and basically any experiment with a new skill or tool.
Designing for feelings, not features
Our v1 was naive and necessary. It asked, “How do you feel?” and tried to make you “happy.”Â
That’s where the first useful truth hit: not everyone wants “happy.” Sometimes you want steady focus for deep work, or calm after a rough commute, or something cozy and handheld when you’re on the go.
We reframed around states like Calm, Focused, Cozy, Energized, and started evaluating every UX decision with one question: Does this nudge the state the user chose?Â
As we tested, we layered control where it mattered: image recognition (scan a dish and get state-aware feedback) and real-time tweaks (adjust spice, carbs, protein, or prep style to steer the outcome).
The interaction is designed to be a simple back-and-forth. You start with how you want to feel, and the AI helps you find or tweak a meal to get you there. It’s not magic and definitely not medical claims.
In parallel, I kept noticing the same pattern everywhere: herbal stores promising calm or energy; friends pairing certain meals with certain moods; caffeine working for some, doing nothing for me. The dots connected late, but when they did, they were obvious:
If we already eat to feel a certain way, why not design for that on purpose?
So we started building PotsandFeelings.com, an AI-augmented meal planner that pairs how you want to feel with meals you’ll actually eat. Not magic and definitely not medical claims. Just evidence-informed choices wrapped in friendly UX.
Evidence-informed guardrails (not prescriptions)
“Psychonutrition” is a rabbit hole; we kept to principles that are practical, common-sense, and adaptable.
We don’t present these as medical protocols. We translate them into guardrails inside the AI so suggestions are aligned with your goal, your diet needs, and your region.
A couple of vivid moments
- The Jollof copy test. We explored how to talk about a familiar dish without flattening it. “Bright spice, steady energy” beat “tasty, balanced carbs + protein” because the first line felt like a promise you could taste, while still pointing to the intended outcome (focus, no crash).
- The “happy” assumption. Early testers reminded us that chasing “happy” isn’t always the goal. Sometimes you want neutral and steady after an anxious morning. That pushed us to reframe the product around state goals, not mood-boosting clichés.
What we promise (and what we don’t)
We promise to help you be more intentional about what you eat relative to how you want to feel using evidence-informed patterns, sensible defaults, and controls you can actually use.
We don’t promise medical outcomes. Pots&Feelings is not a replacement for professional advice. It’s a practical nudge engine for everyday life.
On data: we ask for the minimum needed to personalize (your chosen states, diet needs, optional region), and we use it to improve suggestions. No drama, no dark patterns.
Why this belongs in my portfolio
This project blends product design, UX writing, data thinking, and systems sense into a single, shippable experience. It’s not just “an AI wrapper”; it’s a point of view:
- Start with outcomes, not features.
- Use evidence as guardrails, not as dogma.
- Design transitions, not just labels.
- Build trust with clear defaults, single actions, humane errors, and language that respects the user.
It’s a tiny product with a serious thesis: feelings can be designed for (ethically) when you put the user in control.
What’s next
We’re tightening the state mappings, expanding image recognition, and refining tweak controls. We’ll keep folding tester feedback into small, weekly UX wins instead of saving it for big, brittle releases.
If you’ve ever thought, “I want to feel calm before this call,” or “I need to feel focused for the next two hours,” or “I just want something cozy that won’t slow me down”, this is for you.
Tell me how you want to feel today 👉 Potsandfeelings.com
—
George – designer, builder, and the friend who cooks.