700+Data Points

Real-time sales

23Marketplaces

100%Normalized Data

99.9%Uptime SLA

ChatGPT connectors

Connect ChatGPT to your Amazon Seller Central data

Add one connector in ChatGPT and it reads 700+ documented Amazon data points through 43 tools. SKU profit, ads, inventory, across 23 marketplaces.

Try for Free

Rated 4.7/5 onTrustpilot

How it works

From fragmented Amazon APIs to ChatGPT, in one flow

Nova absorbs SP-API, Ads, and Reports complexity. ChatGPT reads one clean schema.

Step 1

Amazon APIs

Fragmented, schema-shifting

SP-APIrate-limited

Ads APIshifting schemas

Reports APIlags 24 to 48h

We absorb the chaos so ChatGPT does not have to.

Step 2

One schema, pre-modeled for AI

Sales & Profit

Net Profit at SKU levelOrdersFeesCOGSSettlementsReturns

Advertising & Discoverability

PPC KeywordsSearch TermsSessionsBuy BoxBSR

Catalog & Inventory

FBA InventoryListingsCatalogueSubscribe & Save

700+

metrics

marketplaces

SKU

level grain

in real time

orders and units

Step 3

ChatGPT

Reads Nova in one query. Returns profit, fees, ads, inventory answers in seconds.

Real-time sales

ChatGPT connectors let the assistant call live systems instead of reading pasted exports. Amazon's own Selling Partner plugin launched on 23 September 2026 for Claude (in beta, US stores) and Amazon Quick, not for ChatGPT, so the practical route here is an MCP server built on the Selling Partner API. Nova is that server: 43 tools over a reconciled model of your seller account, added to ChatGPT once and available in every conversation after that.

What ChatGPT answers once it reads Nova

Real prompts from real seller teams. Real answers in seconds.

Which 10 SKUs lost the most contribution margin in the last 30 days?

Ranked list with margin delta, fee breakdown, refund rate and ad waste per ASIN. One query against the joined Nova model.

Reads from: Orders × Fees × Ads × Refunds

Why did my profit drop in the UK last week?

Cross-references revenue, ad spend, refunds, FX and fee changes by marketplace. Surfaces the actual cause, not the symptom.

Reads from: Settlements × Ads × Returns

Build a weekly P&L digest with the 5 biggest swings.

Written brief grounded in actual rows: revenue, ad spend, fees, refunds, net profit, with top movers cited.

Reads from: Orders × Settlements × Ads

Which SKUs will stock out in 21 days at current velocity?

Joins FBA inventory, inbound, velocity and lead time. Returns ASIN, days of cover, recommended PO quantity.

Reads from: FBA Inventory × Orders

Find SKUs where ad spend is up 30% but units are flat or down.

Per-SKU ad-spend delta vs unit delta, sorted by waste. Names the campaigns and search terms driving the spend.

Reads from: Ads × Search Terms × Orders

Want more than five? The prompt library has 32, grouped by profit, PPC, keywords, inventory and listings. They run the same in ChatGPT.

query_reviews: query_reviews returns reviews and ratings per ASIN and marketplace on Seller Central accounts: review counts are available only in North America (US, CA, MX, BR), while ratings are available on every marketplace.

query_subscribe_and_save: query_subscribe_and_save returns Subscribe & Save programme health, offers and each product's state from Amazon's replenishment data.

The rest of the Amazon MCP guide

Every page in this cluster answers a different question. Start at the pillar if you want the full picture.

Why ChatGPT on Nova beats ChatGPT on raw SP-API

Same prompt. Nova returns an answer. Raw plumbing returns an error and a token bill.

100+ Amazon sources, pre-modeled

Orders, ads, fees, settlements, inventory, organic. Joined, deduped, ready for one query.

SKU-level profit, to the cent

200+ fee types reconciled per SKU. Net margin, not gross sales dressed up as margin.

Real-time sales, always fresh

Orders and units reach Nova within seconds of the sale, with sales estimated until Amazon confirms the exact amounts. Ad spend refreshes every two hours, search term data three times a day, Amazon finance transactions hourly, and the remaining reports daily.

23 marketplaces, one model

Consolidated or per-marketplace, same schema. No per-region rebuild.

See the full head-to-head with raw MCP

Live in 24 hours, not 24 weeks

Three steps from Amazon to ChatGPT.

Step 01

Connect Amazon

Auth your Seller Central and Ads accounts in Nova. 23 marketplaces supported.

Step 02

Point ChatGPT at Nova

Drop the Nova endpoint into ChatGPT's context. No schema authoring on your side.

Step 03

Ask anything

ChatGPT reads SKU profit, ads, inventory and answers from one joined model.

Also available for

Claude Gemini Nova MCP server 32 prompts to run first Data Warehouses pillar

Questions about Amazon data for ChatGPT

Everything teams ask before plugging Nova into their AI stack

Authorise Seller Central and Amazon Ads in Nova, then add the Nova MCP endpoint as a connector in ChatGPT. From then on ChatGPT can call the Nova tools in any conversation. Amazon's own Selling Partner plugin, launched 23 September 2026, is for Amazon Quick and Claude (beta, US stores), not ChatGPT, so an MCP server over the Selling Partner API is the route.

No. Nova is a remote MCP server, so any MCP-capable client works: ChatGPT connectors, custom GPTs, your own agent, or another assistant entirely. Nova does not lock you into one surface.

Yes. Nova reconciles 370 Amazon transaction lines into 154 P&L lines per SKU: Amazon fees, FBA fees, ad spend, refunds, COGS, VAT. ChatGPT answers in actual margin, not gross sales.

Nova MCP reads by default. Listing edits are available only when a workspace owner turns them on, and every edit is previewed and confirmed before it reaches Amazon. Prices and stock levels are never changed, and Nova does not change campaigns or budgets.

All 23 Amazon marketplaces Nova supports. ChatGPT answers consolidated or per-marketplace questions out of the same model.

Connect your Amazon accounts in Nova, point your ChatGPT client at the Nova endpoint, start asking questions. Most teams are live in under 24 hours.

A raw wrapper makes the assistant discover, join and reconcile the data inside the prompt, and it never returns profit. Nova ships 700+ metrics on Seller Central accounts and 700+ documented data points already modelled, so one call returns the answer. Vendor Central accounts expose additional 1P metrics.

Feed your Amazon data to ChatGPT

Live in 24 hours. Profit-accurate. 23 marketplaces.

Try for Free

Connect ChatGPT to your Amazon Seller Central data

From fragmented Amazon APIs to ChatGPT, in one flow

Amazon APIs

One schema, pre-modeled for AI

ChatGPT

What ChatGPT answers once it reads Nova

The rest of the Amazon MCP guide

Why ChatGPT on Nova beats ChatGPT on raw SP-API

100+ Amazon sources, pre-modeled

SKU-level profit, to the cent

Real-time sales, always fresh

23 marketplaces, one model

Live in 24 hours, not 24 weeks

Connect Amazon

Point ChatGPT at Nova

Ask anything

Questions about Amazon data for ChatGPT

How do I connect ChatGPT to Amazon Seller Central?

Do I need to build a custom GPT?

How fresh is the data ChatGPT sees?

Does ChatGPT get SKU-level profit?

What can ChatGPT change through Nova?

Which marketplaces are covered?

What does setup look like?

How is this different from giving ChatGPT raw SP-API access?

Feed your Amazon data to ChatGPT

Questions about Amazon data for ChatGPT

How do I connect ChatGPT to Amazon Seller Central?

Do I need to build a custom GPT?

How fresh is the data ChatGPT sees?

Does ChatGPT get SKU-level profit?

What can ChatGPT change through Nova?

Which marketplaces are covered?

What does setup look like?

How is this different from giving ChatGPT raw SP-API access?