Upgrade your AI from chatbot to your e‑commerce growth analyst.
The Reaktion MCP feeds your exact e‑commerce metrics directly into Claude, ChatGPT and Grok. Unlock deep analysis, strategic recommendations, and precise campaign optimizations based on data your AI previously couldn't access.
All 1,204 metrics, with their exact formulas.
The complete metric catalog of the Reaktion MCP, pulled live. Tokens show the basis of each one: VAT treatment, refund treatment and conversion source.
All 68 ways to break down your data.
Group any compatible metric by shop, campaign, product, customer segment, channel or time. Each dimension carries descriptive attributes you can add as columns.
467 descriptive attributes.
Names, IDs, prices, dates, URLs and settings that belong to a dimension. Add them as columns next to your metrics, for example the ad name and thumbnail beside spend and POAS.
What you can ask about, and from where.
Every report uses one main period and optional comparisons. Data comes from the sources connected to your Reaktion organisation.
Main periods
Relative periods resolve at the time of the report in your timezone. "Last 30 days" excludes today.
Comparisons
Add to any main period to get a change column.
Custom ranges
Any start and end date. A date-only end includes that whole day; a date-time end is exclusive.
Data sources
Platforms seen connected to Reaktion organisations today. Reports combine them with normalized definitions.
Copy a prompt. Sell more tomorrow.
Tested prompts with the dimensions, metrics and data each one needs. Paste into ChatGPT or Claude and name your store or client.
From question to profit insight in one chat.
No dashboards to build and no exports to stitch together. The MCP plans the report, runs it on your connected data and explains every number.
Connect Reaktion
Add Reaktion as a connector in ChatGPT or Claude and sign in. You only see organisations your Reaktion user can already access.
https://api.reaktion.com/mcpAsk in plain language
Name the store or client, the period and what you want to decide. Use the prompt library when you want a proven starting point.
Powered by Reaktion MCPGet numbers you can trust
Every answer states its basis: VAT, refunds, attribution source and currency. You know exactly what you are comparing.
Under the hood
Four reporting tools run in order every time you ask a performance question.
inspect_reportingFinds the right metric, dimension and period.prepare_reportBuilds and checks a reporting plan.export_reportExecutes the plan on your data.read_reportReturns data with caveats and metadata.Start with the questions your job asks.
Pick your team to see the roles we designed the MCP around and the prompts each one uses most.
Five rules behind every trustworthy answer.
The MCP follows these rules. Use them too when you read, compare or share results with a client or your CFO.
Store financials: excluding VAT, after refunds
Sales, gross profit and net profit are reported without VAT and after refunds unless you ask otherwise. Say it in your prompt to be explicit.
Paid-channel ROAS and POAS use a different basis
Primary ROAS and POAS come from attributed conversions, which can include VAT and are before refunds. They are labelled so you never compare them with store profit by mistake.
Never add or average ratios
Percentages, shares, ROAS and POAS are recalculated from their parts. Summing or averaging them across rows or days gives the wrong number.
Be explicit about time
Refunds land after the order, and each source reports in its own timezone. Compare completed periods and matching weekdays, and expect the last few days to keep updating.
Platforms overlap. Don't sum their conversions.
Google and Meta can both claim the same order. Use store-level metrics like Business-wise POAS when you need one total across channels.
Read-only and permission-aware
The MCP reads data your Reaktion user can access. Agencies see each client separately, and client data is never merged.
Three different states of data
A metric being in the catalog does not mean your organisation has it, and having it does not mean every value is reliable. Good answers say which state they are in.
inspect_reportingThe definition exists in Reaktion. 1,204 metrics today.organisation + grainYour connected sources support this metric at this breakdown.read_report metadataData exists for the period, with caveats stated in the export.Common questions
Does the MCP expose customer personal data?
No. Customer emails, names, phone numbers and addresses are not available through the MCP. Customer-level analysis works on anonymous IDs and aggregated attributes such as market, segment and consent state.
What is the Reaktion MCP?
MCP (Model Context Protocol) is the open standard ChatGPT and Claude use to connect to outside tools. The Reaktion MCP gives them read access to your normalized ecommerce, marketing, analytics and messaging data, with Reaktion's profit definitions applied.
Which data sources are included?
Everything connected to your Reaktion organisation: ecommerce platforms such as Shopify, ad platforms such as Google Ads and Meta, GA4, Klaviyo and imported or affiliate sources. Connection status does not certify freshness, so the MCP reports what it actually found.
Can I report across several stores or clients?
Yes. Reports can span multiple organisations you have access to, converted to each organisation's reporting currency. Agency reports keep each client separate.
Can the MCP change my campaigns?
No. It is read-only. It can recommend budget moves or negatives, but you make the changes in the ad platform.
Why does my ROAS differ from the ad platform?
Reaktion uses your organisation's primary conversion setup, which can include server-side tracked orders the pixel missed. Tracking Uplift shows how many extra orders Reaktion recovers.
Why is net profit missing?
Net profit needs cost settings: product costs, shipping, payment fees and custom cash flows. Without them the MCP labels the metric as unavailable instead of guessing.
Can I build custom formulas or filter rows?
Not yet. Custom formula authoring and account or row filters are unsupported. Use dimensions to break results down and ask the assistant to focus on the rows you care about.
Put every client and product to work on profit.
Connect Reaktion to the AI you already use and start asking.
https://api.reaktion.com/mcp
