Think of Meshi AI as “AI for food data” that occasionally leaks into your life as a chat assistant, not the other way around.
Where I slightly disagree with @sterrenkijker is on how “niche” it is. It starts with menus, sure, but what they’re really building is a structured food graph. That can power a lot more than just “what should I order”:
-
For restaurants / groups
- It becomes a central food database: items, ingredients, tags, historical changes.
- Can plug into POS, online ordering, delivery platforms, etc.
- Useful when you manage many locations or “virtual brands” and don’t want 20 different menu versions to maintain manually.
- Also helps with semi-boring but important stuff: consistency across languages, seasonal menu rollouts, A/B testing descriptions.
-
For apps / platforms
- It’s basically “drop-in intelligence” for food search and recs.
- Instead of building your own NLP and tagging around menus, you just feed them data and hook into their APIs / widgets.
- This matters if you’re building a food delivery app, a nutrition app, or even corporate cafeteria software.
-
For consumers
- The chat UX is more a thin layer on top of that data.
- When it works, you get: “I’m low FODMAP, training tomorrow, give me high protein and not too greasy” and it surfaces dishes that roughly match.
- But you’re still limited by what the restaurant actually provides and what’s integrated. It’s not a magic nutritionist in your pocket.
What it really solves in practice:
- Fragmented, inconsistent menu data scattered across PDFs, POS exports, Excel files.
- The pain of adding nutrition / allergen tags at scale.
- Crappy search like “vegan” returning 4 items when you actually have 20 that would qualify if someone tagged them right.
Where I’d be cautious:
- If you care about strict medical nutrition or allergies, treat outputs as filters and hints, not gospel. AI infers ingredients from names, and that’s inherently risky.
- Vendor lock-in: once all your menu intelligence, rec rules, and tags live in Meshi, migrating out will hurt. Export options & contracts matter.
Is it “worth it” for you:
-
You’re a restaurant / chain / dark kitchen:
- Worth looking at if menu complexity or data entry is eating staff time.
- Less compelling if you have a tiny static menu and already nailed allergen / nutrition workflows.
-
You’re just a diner:
- It’s only worth caring about if it’s quietly baked into apps you already use. You probably won’t get huge value as a standalone “I signed up for Meshi” thing.
- If decision fatigue is a big problem for you or you have multiple dietary constraints, it can be nice when it’s available, but I wouldn’t reorganize my life around it.
TL;DR:
Not a general-purpose AI buddy, not just a single app either. It’s more like an infrastructure layer that makes food data usable, with a chat interface slapped on top so humans can interact with it in normal language.