Introducing Remi, by Basket Labs

The AI assistant that builds your shoppers’ meals and cart, on your catalog.

Remi is an end-to-end, white-label AI assistant from Basket Labs. It runs on our proprietary data infrastructure, is embedded directly in your storefront and app, and tailors every meal and cart to the shopper’s nutrition targets, budget and tastes, using the products and prices in your catalog.

A Remi recipe page showing a dinner with its nutrition figures and adjustable ingredientsA Remi cart showing meal cost, checkout cost and the product selected for an ingredient

What Remi can do

Remi takes calorie, protein and budget targets as numbers, alongside diet, allergies, tastes and cooking style, and uses them to build recipes and select SKUs.

Remi screen where a shopper sets calorie and protein targets for each meal

Every Remi capability runs on the same data and AI infrastructure. That’s why we can customize it for your business and ship new features fast.

The four types of shoppers Remi brings to your e-commerce platform

The nutrition-minded shopper - Robert

Robert is on a mission to lose weight and gain muscle. He’s on a GLP-1 and in the gym every day, but hitting 1,600 calories and 100 grams of protein a day is hard, and nutrition tracking apps only allow him to log meals after the fact. Remi helps Robert build a plan that hits his daily numbers, tastes good, and puts every ingredient in his cart.

27%of US adults have tracked their diet or calories with an app.

The AI adopter - Maya

Maya asks ChatGPT before she decides anything, including how to make gluten-free pescatarian dinners, which grocer to shop at, what to shop for at the store, and how to cook her meals. But ChatGPT can’t see her store’s shelves, and it only helps with one task at a time. Remi helps Maya plan her dinners, fill her cart from her store’s actual shelves, and cook each meal, all in one place.

70%of US consumers use AI to shop.

The meal-kit shopper - Dana and Chris

Dana and Chris get a meal-kit box every week. They like not having to think about dinner, but the menu is limited and it costs more than groceries. With Remi, they pick the dishes they want, get right-sized ingredients from their own store, and cook with step-by-step help, at a fraction of the price.

21%of adults under 45 eat from a meal kit in a typical week.

The budget shopper - Angela

Angela struggles to feed a family of four on $250 a week. She plans around the flyer, clips coupons and adds it up in the aisle. With Remi, she sets her budget, and the week’s meals and cart are built to it, around her family’s tastes, at the store’s actual prices.

88%of Americans have changed how they grocery shop to save money.

How Remi compares to grocery AI currently on the market

Remi’s AI ExperienceInstacartClementineCooklistDeployed with Kroger bannersGrocer-built assistantsAlbertsons AIDoorDash / Uber Eats
OwnershipGrocer-owned, white-label
Runs in the grocer’s own storefront, grocer keeps the shopper data
Runs on Instacart’s platform
White-label, deployed with Kroger
Built in-house
Marketplace only
PlanningDecision dimensions covered
22 structured, plus free text
A few at a time via chat, poor adherence
Complete meal plans
Unified creation with full slot customization across days/meals
Recipes treated as independent items
Recipe-level editing
Recipes can be edited at ingredient level
GoalsQuantitative calorie/protein goal adherence
Targets hit through recipe construction and product selection
Approximated, mixed adherence
Accurate but limited to recipe library
Weekly budget adherence
Week built to a dollar budget at the grocer’s prices
Recipes include expected cost
ShoppingTailored SKU recommendations
Picked for recipe, budget, nutrition goals
Must be prompted with a specific request
Consolidated cart
Built-in consolidation logic
CookingPer-recipe cooking assistant
Cooking assistant tailored to each recipe

As of November 2026

The best grocery AI only works on the best infrastructure. So we built our own.

Data infrastructure built for grocery AI

  • Your catalog gets stored as concepts with their relationships, not as strings: what each product is, what it’s made of, what can stand in for it. That turns your catalog into something LLMs can reason over, with substitution, package math and quantitative logic built in.
  • It covers every SKU in every store, with price and availability. Remi reads the catalog while it works, so it avoids hallucinations and recipes that don’t match inventory.

AI infrastructure that’s a product, not a system prompt

  • Remi runs on a pipeline of specialized AI modules, each built for a purpose.
  • Dietary restrictions, nutrition limits, meal prep and substitutions are engineered in as core behavior, not left to a general model as edge cases.
  • The build is modular, so the same pipeline runs as a planner, a chat assistant or a search tool.

AI products have to make compromises between depth, accuracy, and cost. Remi optimized for all three.

22Decision dimensions covered
<5%Error rate across decision dimensions
99%+Of ingredients matched to real SKUs in the grocer’s catalog
$0.01End-to-end token cost per meal

A customer acquisition and retention engine for your e-commerce.

An AI layer is forming between shoppers and grocers, and the habits being set now will stick. Remi helps you own that layer on your own platform, instead of ceding it to third parties.

Habit and loyalty

Remi is built for frequency. The cooking assistant brings shoppers back to your app every day they cook, and planning the next shop brings them back every week. Every session happens in your storefront, on your brand.

Bigger, better baskets

Shoppers who plan with Remi buy the whole basket at your store instead of splitting it across two or three. And because Remi weighs price and value on your real catalog, private label surfaces naturally.

Capturing the right shopper

Remi’s natural audience is the online-first, AI-comfortable shopper, the customer most likely to drift to a competitor’s app. Giving them a reason to plan, shop and cook in your storefront is both defense and acquisition.

First-party shopper data

With no middleman, Remi can offer better personalization based on what your loyalty program already knows about the shopper, and it opens up opportunities for direct retail media revenue. It also generates new data on how your shoppers plan and cook.

What a partnership with Basket Labs looks like

Deploy now, not later

In the AI adoption race, every day matters. Remi is ready today, and its modular design means fast deployment on your platform.

We do the heavy lifting

We work with you to embed Remi into your e-commerce platform, then help manage everything behind it.

Built to fit you

Remi is highly customizable: which features to include, what to surface, how it looks. It ships in your brand, and our flexible architecture makes grocer-specific requests easy to accommodate.

AI without the R&D

As AI keeps evolving, our team maintains and improves Remi: adjusting to new models, increasing efficiency and developing new capabilities. Through our data and AI infrastructure, we can design and build your AI ideas in a fraction of the time. That includes agentic shopping: Remi was designed for it from the start, so it’s there when you want to offer it.

See Remi running on your own catalog.

Or email info@basketlabs.ai