AI demand forecasting for Shopify

Order the right amount. Prove it saves money.

NeuraOrder forecasts every product with an AI model, shows the odds you'll run out before your next delivery, and turns that into purchase orders sized to your budget, shelf space and supplier terms.

Launching on the Shopify App Store. Early-access members get the 30-day free trial first.

Example: ceramic pour-overDaily sales · next 21 days
Today Likely range Order arrives, day 7
120 unitsSuggested order
5%Chance of running out (64% without it)
$1,480Sales protected
Tested on real sales, not promises

The research behind every order

In a replay of 26 weeks of real sales from a 1,000-product online store, NeuraOrder cut the cost of lost sales and carrying stock by $91,400, or 48%, compared with a common reorder rule.1

Where the money went

Cost of lost sales plus carrying stock, 1,000 products over 26 weeks

  • Sales lost to stockouts
  • Cost of carrying stock
Common reorder rule Common reorder rule: sales lost to stockouts, $174,150 Common reorder rule: cost of carrying stock, $15,026 $189.2k $174.2k lost sales + $15.0k carrying stock NeuraOrder NeuraOrder: sales lost to stockouts, $63,781 NeuraOrder: cost of carrying stock, $33,947 $97.7k $63.8k lost sales + $33.9k carrying stock

NeuraOrder spent more on carrying stock (it held 50 days of supply, the rule 22) and far less on lost sales.

Data: UCI Online Retail II (Chen 2012, CC BY 4.0), replayed through a simulated stockroom. Assumptions in note 1; the full method is on the research page.

77%

Fewer weeks out of stock1

Stock ran out in 2.9% of weeks, against 12.6% with the common reorder rule.

93.7%

Of customer demand served1

Up from 82.0% with the reorder rule: 11.7 more points of sales captured.

40%

Less fresh-food waste2

When NeuraOrder knows each product's shelf life, across two samples of 400 grocery products. Profit rose 5–7%.

3–12%

From estimate to real savings3

Bulk-buy savings came in within 3–12% of what NeuraOrder predicted before the order.

  1. Replay of real daily sales for 1,000 products of a UK online retailer (UCI Online Retail II, Chen 2012, CC BY 4.0) through a simulated stockroom: 14-day supplier lead time, weekly ordering, unmet demand lost, 50% margin and 25% yearly holding cost assumed. The "common reorder rule" orders up to 1.5× the last 28 days' average sales for the next 21 days. NeuraOrder held more stock where running out cost more than carrying it (50 vs 22 days of supply).
  2. FreshRetailNet-50K (Wang et al. 2025, CC BY 4.0), two independent samples of 400 products, simulated stockroom with a 6-day shelf life, compared with NeuraOrder ordering without the shelf life set. Waste fell from 11.6% to 7.0% and from 8.3% to 4.7% of units bought.
  3. Real UCI sales with simulated supplier volume discounts and announced price increases, at NeuraOrder's default (cautious) bulk-buy setting.

Results come from simulations on public datasets, not from customer stores, and your results will depend on your products, margins and suppliers. Read exactly how we tested.

What you get

Every reorder decision, with the odds and the money in plain sight

Odds, not guesses

A likely range for every product

See the chance of running out before your next delivery, and pick the in-stock level you want. Orders are sized to hit it.

Proof on your store

Check it on your own sales

Before you trust a forecast, the Forecast accuracy page shows how it would have done on your store's past sales.

Money view

A dollar figure on every order

Each suggested order shows the sales it protects and what skipping it is likely to cost, with a range.

Bulk buys

Buy ahead only when it pays

Supplier price rises and volume discounts weighed against interest, storage and the risk of slow sales, at the appetite you choose.

Your limits

Storage space and shelf life

Orders that fit your shelves when they arrive, and perishables sized to sell before they expire.

Purchase orders

From suggestion to supplier

Purchase orders by supplier as PDF, CSV or email, tracked as incoming stock until they arrive.

How it works

Set up in minutes, forecasts every night

  1. Install from Shopify

    NeuraOrder reads your products, inventory and order history. It never changes your store.

  2. Tell it how you buy

    Supplier lead times, how often you reorder, minimum orders and case packs. Store-wide defaults take a minute.

  3. Reorder with confidence

    Every night you get fresh forecasts, a reorder plan and stockout alerts. Turn them into purchase orders you send yourself.

Your data stays yours

Private by design

NeuraOrder never collects your customers' personal information. Your data is never sold and never used to train AI models. Contact details you enter are encrypted, and everything is deleted 48 hours after you uninstall.

Switching from Stocky?

Shopify retired Stocky on August 31, 2026. NeuraOrder brings back forecasting and purchase orders, with AI forecasts and proof on your own sales.

How to switch
Pricing

From $19 a month, every feature included

Plans are based on how many products (SKUs) you forecast. Every plan starts with a 30-day free trial, billed through your Shopify account.

Compare plans
Questions

Good to know

Is it really AI forecasting?

Yes. NeuraOrder combines a pretrained AI forecasting model with proven statistical methods. It was tuned on one set of products and checked on a separate set it had never seen, so its likely ranges are honest. It never trains on your data.

How soon will I see forecasts?

Right after setup for products with about 8 weeks of sales. Newer products are forecast as soon as they have enough history.

Does NeuraOrder change my store or place orders?

No. It only reads your store data. Purchase orders are drafts you review and send to your suppliers yourself.

How does the free trial work?

Every plan starts with 30 days free. After that, the plan is billed through your Shopify account. Uninstall before the trial ends and you pay nothing.

Can I trust the research numbers?

They come from replaying real, public retail sales through a simulated stockroom, with every assumption stated. The research page shows the method, the full results and the limits. Inside the app, you can check accuracy on your own sales.