Nova Analytics

AI Agents for your Amazon business

Ship agents that watch margin, plan inventory and brief your team every Monday.

Step 1
Amazon

Amazon APIs

Fragmented, schema-shifting

SP-API
Ads API
Reports API
Nova
Step 2

All your Amazon data, one schema

Sales & Profit

Net Profit at SKU levelOrdersFeesCOGSSettlementsReturns

Advertising & Discoverability

PPC Keywords & CampaignsSearch TermsSessionsBuy BoxBSR

Catalog & Inventory

FBA InventoryListingsCatalogueSubscribe & Save
200+
metrics
21
markets
SKU
grain
Hourly
refresh
Custom Amazon Agents

AI Agents for your Amazon business

Example agents

Profit CheckerRestock CopilotAd Spend AuditorACOS BreachCustom P&LBusiness TargetsMonday Brief
ClaudeClaude
ChatGPTChatGPT
GeminiGemini
Try for Free
Rated 4.9/5 · 14-day free trial

Agents you can ship in days

Profit CheckerRestock CopilotAd Spend AuditorACOS BreachCustom P&LBusiness TargetsMonday Brief
SellerX
Big Fudge
Bosike
Sun Ninja
Morabella
Portstrong
GBL
Growster
SellerX
Big Fudge
Bosike
Sun Ninja
Morabella
Portstrong
GBL
Growster
What teams ship on Nova

Build anything your Amazon business needs

Powered by the most accurate dataset in the Amazon ecosystem.

Profit Checker

Daily SKU profit guard

Breaks down every SKU's profit to the cent and pings you when one starts bleeding.

  • Revenue, fees, ads and COGS on one waterfall
  • Names the line item killing the margin

Reads · Orders, Fees, Settlements, Ads

Runs on
ClaudeChatGPTGemini
Build this agent
● Live on Nova data
SKU-B0317 · last 7 days

Net profit · SKU-B0317

$0−18% WoW
Revenue$12,840
-Amazon fees$3,184
-Ads spend$2,491
-COGS$4,729
Profit alert · storage fee +38% after Q4 surcharge on SKU-B0317

How it works

From your seller data to an agent built around your team.

Hourly Amazon data on one side, your AI agent on the other. Here is what happens in between.

01

Connect Amazon APIs to Nova

Give us access to your data in a few clicks.

Seller CentralAdvertising
Nova
Connect API
02

We provision your database

We spin up your dedicated ClickHouse database with host, ID and password, ready to query.

NovaClickHousehost · ID · password
Handled by Nova
03

Plug into Claude, ChatGPT or Gemini

Drop your credentials into your AI assistant of choice so it answers with full context of your Nova data.

host · ID · password
ClaudeChatGPTGemini
See examples
04

Prompt your way in

Prompt your agent for daily margin checks, restock plans, ad audits, the Monday brief.

Your promptAgent answer
See agents you can ship

Bring your own Claude, ChatGPT or Gemini account. Nova handles the data, your assistant handles the answers.

Hourly refresh. 21 marketplaces.
Live connection, zero sync

How the connection works, and why it's super powerful

Four secrets. One pipe. Your ClickHouse stays where it is, your agent stays yours, and the data is live, not synced.

Your warehouse
Your ClickHouse
Self-hosted or ClickHouse Cloud
USUKDEFRJP+16
Add secrets
CLICKHOUSE_HOST
xxxxx.eu-central-1.aws.clickhouse.cloud
CLICKHOUSE_USER
nova_share_user
CLICKHOUSE_PASSWORD
••••••••••••••••
CLICKHOUSE_DB
nova_share
Live in your agent
What SKUs lost margin yesterday?
SKU-B0317-4.2ptFBA fee
SKU-B0192-2.8ptAd spend
SKU-B0405-1.9ptReturns
ClaudeChatGPTGemini
Your AI assistant
Answers query ClickHouse in real time
Zero data movement
Your warehouse stays yours. Your assistant reads, never copies.
Live queries, no cache
Every answer hits ClickHouse in milliseconds.
One source of truth
Same numbers in Nova, your BI tool, and every agent answer.
Switch off anytime
Revoke the credential, the agent goes silent. No vendor lock.

Five reasons

Why Nova AI Agents are unique and unmatched

01Built for Agents

The foundation your custom Amazon Agents need.

Pre-modeled schema, sub-second responses, zero token-burn from schema discovery. Claude, ChatGPT and Gemini get the answer in one shot, so your agents act on real data instead of guessing.

See why this is much more powerful than an MCP to actually build workflows
Nova · pre-modeled layer
Claude logoClaude
ChatGPT logoChatGPT
Lovable logoLovable
Gemini logoGemini

Same pre-modeled data. Pick your AI Assistant.

02Data coverage

Every Amazon metric. In one place.

Exhaustive coverage across sales, finance, inventory, ads, catalog and search intelligence. Missing a metric? We add it within 48 hours, on demand.

Modeled domains

Live in production
  • Sales & Orders5/6
  • Finance & Profit5/6
  • Inventory & FBA5/6
  • PPC & Advertising4/6
  • Catalog & Listings5/6
  • Customer & Search Intelligence4/6

100+ SP-API and Ads-API endpoints, one stable schema.

03Profit to the cent

The most accurate dataset on the market.

One year building the most accurate Amazon P&L on the market. SKU-level, settlement-reconciled, every fee accounted for. The line between a nice gadget and a system finance trusts.

300+ rules appliedPer SKU · Per day
Gross revenue$128,420.18
−Amazon fees (FBA, referral, storage)$31,847.06
−Ads spend (SP, SB, SD)$24,910.55
−COGS & inbound$40,228.92
=Net profit (to the cent)$31,433.65

Generic providers stop at aggregated totals.

04Hourly refresh

Data refreshed hourly.

Nova refreshes every hour. Competitors deliver every 24 hours. You see Amazon moves before they do.

Refresh cadence

T-1h · live
  1. 08:00
    synced
  2. 09:00
    synced
  3. 10:00
    synced
  4. 11:00
    synced
  5. 12:00
    synced
  6. 13:00
    synced
  7. 14:00
    now

21 marketplaces, every hour. Same freshness in dashboard, warehouse and agent.

05Data Enrichment

Enrich data to provide accurate context to your agents.

Tag products, add COGS, FBM fees and VAT directly inside Nova. Enrichment flows into every query, dashboard and agent automatically.

Smart tagging on Nova SKUs
COGS FIFO and period-based editor
COGS · FIFOTagsFBM costVAT

MCP vs Nova AI Agents

A raw MCP gives your agent a socket. Nova gives it answers.

Same Claude, same prompt. Nova-backed agents return reconciled answers. Raw MCP returns an error and a token bill.

Dimension
Query cost
Raw MCP8 to 20 tool calls per question
Nova1 SQL query, full answer
Schema stability
Raw MCPBreaks when SP-API changes
NovaPre-modeled and version-stable
Profit logic
Raw MCPNone. LLM invents COGS and fees
NovaNet profit to the cent per SKU
Freshness
Raw MCPRate limits leak into chat
NovaHourly refresh, throttling absorbed
Coverage
Raw MCPOne report at a time
Nova100+ sources, 21 marketplaces

Your AI should answer in profit and SKUs, not in SP-API endpoint names.

Connect API

Dedicated pages per AI assistant

Amazon Data for ClaudeAmazon Data for ChatGPTAmazon Data for GeminiAmazon MCP for Sellers

Frequently Asked Questions

Everything teams ask before plugging Nova into their stack

The Amazon data layer for AI Assistants. 100+ SP-API and Ads-API sources pre-modeled, SKU-level profit reconciled across 200+ settlement items, refreshed hourly. Pipe it into your warehouse, your agent, or your custom app on day one.
Most teams underestimate the work: 6 to 12 months to model the schema, more to reconcile profit, and a permanent maintenance burden every time Amazon ships a breaking change. Nova absorbs all of that. You start querying real, profit-accurate data in 24 hours.
Reporting Automation is one CSV link for Sheets, Excel, or your existing BI dashboards. Data Warehouses is raw, SKU-level, intraday data inside your warehouse, agent context, or custom app. Same source of truth, different surface.
Yes. The schema is stable, versioned and agent-ready. Teams ship Claude, ChatGPT, Lovable and custom in-house agents on top of Nova in days instead of months. No vector store gymnastics, no SP-API plumbing.
100+ sources covering sales, orders, finances (200+ settlement items), FBA inventory, Sponsored Products/Brands/Display at the product level, listings, catalog, pricing, Buy Box, SQP, Brand Analytics search terms, BSR on your ASINs, reviews. Refreshed hourly across 21 marketplaces.
Warehouses: Snowflake, BigQuery, Redshift, Databricks, ClickHouse, Postgres, MySQL. BI: Looker Studio, Power BI, Tableau, Hex, Metabase. Standard MySQL wire protocol means most tools connect natively.
Sources flagged 'In progress' are actively being modeled and have a target date. 'Roadmap' is on the plan with no committed date yet. Everything marked 'Live' is in production and available the day you connect.
Most teams query live Nova data within 24 hours. Connect your Amazon accounts, pick your destination (warehouse, MySQL endpoint, or agent), and you are live. Profit data hydrates within the first refresh cycle.
Already in the model. Upload COGS with effective dates, tag and group SKUs into portfolios, brands or categories inside Nova. No external PIM to wire. The shape your team works in is the shape that lands in your warehouse.
Book a demo. We scope your destination, marketplaces and use case, then quote a plan. Most customers are live within a week.