§0 AI-native reporting for ecommerce operators

Turn scattered ecommerce data into weekly reports, creative insight, and next actions.

Kreel connects Shopify, Meta Ads, Google Ads, and Klaviyo, normalizes the data, and gives your AI agent the context to analyze performance, explain what changed, and produce client-ready reports.

Request beta access See example workflows

Run

/kreel:weekly-report demo-brand

Reads

Shopify + Meta + Google + Klaviyo

Returns

Analysis, creative insight, reports, and next actions

Connects to

Shopify

Klaviyo

Meta Ads

Google Ads

Works in

Claude DesktopClaude CodeClaude.ai ChatGPT^*

*via custom GPT until OpenAI ships MCP

§1 The problem

Ecommerce reporting still starts with too much manual work.

Your numbers live in Shopify, Meta Ads, Google Ads, and Klaviyo. To explain a single week, someone exports four dashboards, screenshots the charts, drops them in a spreadsheet, and reconciles the figures by hand before any thinking starts. Every report begins from zero.

Today · scattered sources

Shopify exportMeta screenshotGoogle CSVKlaviyo dashboardSpreadsheetSlides

With Kreel

One AI-ready context

§2 What you get

Built around the outcomes, not the API.

Kreel gives your AI the operating context behind your store, already normalized and ready to reason with.

Unified performance data

Shopify, Meta, Google, and Klaviyo normalized into one model. Revenue, spend, orders, campaigns, flows, and cohorts line up on the same dates and currency.

AI-ready context

Data shaped for a language model: compact, labeled, consistent. Your agent reasons instead of parsing raw API dumps.

Creative and campaign intelligence

Creatives, campaigns, and keywords with the metrics to explain what's working, what's fatiguing, and what to brief next.

Reusable operator workflows

Battle-tested report and analysis skills you run in the AI you already use. Recurring reviews take one prompt.

§3 Workflows

The reviews you run every week, as one prompt.

Weekly performance report

Revenue, spend, blended efficiency, and what changed, with next actions.

Shopify · Meta · Google · Klaviyo
Creative performance review

Top and fatiguing creatives by CTR, hook rate, and CPA, with what to brief next.

Meta · Google
Retention and Klaviyo review

Flow and campaign performance, AOV lift, and cohort retention.

Klaviyo · Shopify
Budget and channel review

Where the next budget should go by marginal efficiency across channels.

Meta · Google · Shopify
Founder/client update

A plain-language monthly summary: results, drivers, and the plan.

All platforms

§4 Example output

Example output: a weekly report you can send.

Reports are one visible output of the data layer. Ask for the weekly view and your agent can return narrative takeaways, source metrics, and next actions from the same cross-platform context.

Example command

kreel run weekly-report --client demo-brand --range last_7d

Weekly report

Demo Brand · sample week

Sample · 1 Kreel call

Revenue

€84.2k +9.2%

Orders

612 +6.8%

MER

3.48x -0.21

Email revenue

€10.0k +14.1%

What changed

Revenue rose while blended efficiency softened. Meta contributed the largest week-over-week swing, Klaviyo flows lifted AOV on the ad-push cohort, and Google stayed efficient but concentrated in brand search.

Shopify total €84,210

Meta Ads attributed €16,892

Google Ads attributed €18,044

Klaviyo flows €10,012

Recommended next actions

1 Move 12% of prospecting budget from retargeting-core into summer-launch.
2 Keep browse-abandon flow live; AOV is €12.40 above the cohort baseline.
3 Review Google brand search overlap before increasing spend next week.

Source context

Shopify orders, Meta campaigns, Google campaigns, Klaviyo flows, customer cohorts.

§5 Creative

See which creatives are working, and which are fatiguing.

Every active creative with its spend, ROAS, revenue, CTR, and CPM, tagged by hook, format, audience, and product. Your agent reads the same view to explain what's driving results and what to brief next.

Kreel creative dashboard: a grid of ad creatives tagged by hook, format, and audience, each showing spend, ROAS, revenue, CTR, and CPM with week-over-week deltas. — Creative dashboard · sample data

Tag & hook breakdown Fatigue signals Spend · ROAS · CTR · CPM By audience, format, product

§6 Try it · live

Ask across the whole account, not one platform.

Routed through a real agent against a demo workspace. The agent gets compact ecommerce context instead of raw dumps from separate tools.

— Demo Brand, last 7d

Try it · kreel + claude live

you

⌘ ↵ to run

kreel:

§7 The data layer

One key. Every platform. Any agent.

One API key, four platforms, every agent. We ingest through OAuth, keep the platform data synced in Kreel, normalise it across clients, and serve it to whatever you already use.

Shopify

orders, customers, products

Klaviyo

flows, campaigns, profiles

Meta Ads

campaigns, creatives, audiences

Google Ads

campaigns, keywords, conversions

your key

kreel_live_8f3e…c02a

Claude Desktop

one-click extension

Claude.ai

connector URL

Claude Code

kreel auth login

ChatGPT

custom GPT

One call. Four platforms. Normalised. EU-hosted · OAuth-only · read-only

§8 Skills

Install Kreel. Inherit our agency's playbook.

Every Skill was battle-tested on real client work before it shipped. Run them in the agent you already use.

See all 8 Skills →

weekly-report ~2 hrs

Full client report, ready to send.

meta-health ~30 min

7-day Meta Ads health check.

anomaly-detective ~1 hr

Spot spend anomalies before your client does.

creative-brief ~45 min

Auto-generated agency creative brief from performance.

crm-calendar-planner ~2 hrs

Interactive email content calendar.

budget-advisor ~1 hr

Where to shift spend this week.

ad-pack-builder ~1 hr

Copy pack from a folder of creatives.

monthly-report ~3 hrs

Tabbed HTML dashboard for client delivery.

§9 Safety

Read-only by design.

Kreel is an analysis and reporting layer first. At launch it reads your data to explain performance and draft reports. It does not edit campaigns, change Klaviyo flows, update Shopify products or discounts, or touch customer data. Connections use scoped, read-only OAuth you can revoke anytime.

No campaign edits
No flow changes
No product or discount writes
No customer-data changes

§10 Why not just connect every MCP?

Better than wiring up four MCPs yourself.

Pulling ecommerce performance through separate tools means asking Meta, Google, Shopify, and Klaviyo for each date range one by one. Kreel does that work in the data layer, then returns the compact answer your agent needs.

Typical MCP stack

meta.get_insights(last_7d)

meta.get_insights(last_30d)

google.get_campaigns(last_7d)

shopify.get_orders(last_7d)

klaviyo.get_flows(last_7d)

More tool calls, repeated date logic, platform-specific field names, and a bloated context window before the agent can reason.

Context window used 86%

Little room left for synthesis, recommendations, or a client-ready report.

Kreel

kreel.get_performance({

client: "demo-brand",

range: "last_7d",

include: ["shopify", "meta", "google", "klaviyo"]

})

1 call All platforms

1 schema Normalised fields

Low tokens Analysis-ready

Context window used 18%

More of the window stays available for analysis, narrative, and next actions.

Shopify orders, customer cohorts, ad spend, campaign performance, and Klaviyo flows arrive together, already shaped for an LLM.

§11 How it works

Three steps. Under five minutes.

1 Step 1

Connect your clients' platforms

OAuth Shopify, Klaviyo, Meta, and Google in minutes. No passwords, read-only by default.
2 Step 2

Get your Kreel key

One key per workspace. Looks like kreel_live_… Scoped to that workspace only.
3 Step 3

Paste it into the AI you already use

Claude Desktop (one-click extension), Claude.ai (connector URL), Claude Code (kreel auth login), or the CLI directly.

§12 For builders

A single schema, already normalised.

Under the report layer is a compact schema for orders, customers, campaigns, flows, and cohorts. Your agent stops guessing at field names and starts answering questions.

+ 14 more endpoints

{
  "range": "2026-04-11..2026-04-17",
  "workspace": "demo-brand",
  "totals": {
    "revenue":     84210.00,
    "orders":           612,
    "aov":          137.60,
    "returning":     41.3
  },
  "channels": {
    "direct":        38210.50,
    "google_ads":    18044.20,
    "meta_ads":      16892.10,
    "klaviyo_flows": 10012.30,
    "other":           1051
  },
  "top_products": [
    { "sku": "SER-01", "rev": 12044, "qty": 86 },
    { "sku": "CRM-02", "rev":  9872, "qty": 74 }
  ]
}

{
  "segments": {
    "new":           341,
    "returning":     271,
    "vip_gte_3":      58
  },
  "cohorts": [
    { "month": "2026-02", "size": 420, "m1_ret": 0.38 },
    { "month": "2026-03", "size": 512, "m1_ret": 0.41 }
  ],
  "aov_by_segment": {
    "new":       124.50,
    "returning": 158.90,
    "vip":       234.10
  }
}

[
  { "id": "23856…01",
    "name": "summer-launch",
    "spend":   4200.12,
    "clicks":   18204,
    "purchases":   301,
    "roas":       4.02,
    "attributed_revenue": 16884.50 },
  { "id": "23856…02",
    "name": "retargeting-core",
    "spend":   2810.00,
    "clicks":   11048,
    "purchases":   172,
    "roas":       2.88,
    "anomaly": "spend +31% vs 7d avg" }
]

[
  { "id": "welcome-01",
    "triggered": 1104,
    "opened":     682,
    "placed_order": 78,
    "revenue":   4210.50 },
  { "id": "browse-abandon",
    "triggered": 2208,
    "opened":   1104,
    "placed_order": 91,
    "revenue":   5802.10,
    "aov_delta_vs_baseline": "+€12.40" }
]

§13 Who it's for

For operators who already know what good ecommerce analysis looks like.

Kreel assumes you can read a P&L and a MER chart. It removes the gathering, not the judgment.

Boutique agencies Fractional CMOs Growth leads Retention consultants Paid media teams AI-native ecommerce operators

§14 Built on our own book

"We built Kreel for our own agency first. We run every client's weekly review through it before it ever ships to anyone else."

Beyond Peaks

360° ecommerce agency · Paris

platforms unified: 4
agent surfaces: 4
hosting: EU
data access: read-only

§15 Where it fits

Not the only way. Often the right one.

	Kreel	Polar Analytics	Ads-only MCPs	Build it yourself
Platforms	Shopify · Klaviyo · Meta · Google	All four +	Ads only	You build it
Primary surface	MCP-first	Dashboard-first	MCP	Your call
Skills library	8 beta skills	No	No	No
Monthly cost	Free demo · €99–499	€500+	Varies	Dev time
Setup to first value	Minutes	Hours–days	Minutes	Weeks

§16 Pricing

Pricing that scales with the number of stores you report on.

Plans for when your beta access opens. Pricing is set; billing starts when you go live on real data.

Operator

€99/mo

Solo operator · 1 store

1 client workspace
All 4 platforms
Report Skills library
Claude, ChatGPT, MCP + CLI

Request beta access

Team

€249/mo

Fractional · 2–3 brands

3 client workspaces
All 4 platforms
All agent surfaces
Priority email support (24h)

Request beta access

Agency

€499/mo

Multi-client agency

10 client workspaces
Extra workspaces at €39/mo
Priority support + onboarding
Shared workspace permissions

Request beta access

Free playground + Scale

Beta access includes a demo workspace. Larger agencies get custom workspace limits, SSO, Slack support, and priority SLAs.

Contact sales

§17 Security

Boring where it should be.

OAuth-only

Never your passwords. Only scoped OAuth tokens you can revoke at any time.
EU-hosted

Primary compute and database in the European Union. Transfers covered by SCCs.
Read-only by default

Kreel reads to analyze and report. It does not edit campaigns, flows, products, discounts, or customer data at launch.

Read the security page →

§18 FAQ

Questions you'll ask.

What is Kreel?

An AI-ready ecommerce context layer. It connects Shopify, Meta Ads, Google Ads, and Klaviyo, normalizes the data, and serves it to the AI you already use through an MCP server, a CLI, a Claude Desktop extension, and a Claude.ai connector.
How do I get access?

Kreel is in private beta. Request beta access and we onboard operators and agencies directly.
Does Kreel change anything in my accounts?

No. Kreel is read-only at launch. It does not edit campaigns, change Klaviyo flows, update Shopify products or discounts, or touch customer data. Connections use scoped OAuth you can revoke anytime.
Do I need separate MCPs for Shopify, Meta, Google, and Klaviyo?

No. You connect those platforms to Kreel once, then your agent calls Kreel for the combined ecommerce context instead of juggling one tool per platform.
Is Google Ads supported?

Yes. Kreel connects Google Ads alongside Shopify, Meta Ads, and Klaviyo, so reporting can include campaign, keyword, conversion, and ecommerce context in the same workflow.
Do I need to be technical?

No. For most users, that means pasting a connector URL into Claude Desktop or Claude.ai — three clicks.
How is this different from Polar Analytics?

Polar is a full dashboard and attribution suite. Kreel is the data layer your agent reads — a fraction of the price, with a curated Skills library.
What about security and data residency?

OAuth-only, EU-hosted, read-only. See the Security page for full details.
Can I manage multiple clients?

Yes — the product is built around agencies managing a book of clients. Each client is fully isolated.
When will TikTok, GA4, and Amazon support land?

On the roadmap. We publish progress in the Changelog.

Bring your ecommerce data into one AI-ready context.

Kreel is currently opening private beta access for ecommerce operators and agencies who want to automate reporting, creative analysis, and recurring performance reviews.

Request beta access

Turn scattered ecommerce data into weekly reports, creative insight, and next actions.

Ecommerce reporting still starts with too much manual work.

Built around the outcomes, not the API.

Unified performance data

AI-ready context

Creative and campaign intelligence

Reusable operator workflows

The reviews you run every week, as one prompt.

Weekly performance report

Creative performance review

Retention and Klaviyo review

Budget and channel review

Founder/client update

Example output: a weekly report you can send.

Demo Brand · sample week

What changed

Recommended next actions

See which creatives are working, and which are fatiguing.

Ask across the whole account, not one platform.

One key. Every platform. Any agent.

Install Kreel. Inherit our agency's playbook.

Read-only by design.

Better than wiring up four MCPs yourself.

Three steps. Under five minutes.

Connect your clients' platforms

Get your Kreel key

Paste it into the AI you already use

A single schema, already normalised.

For operators who already know what good ecommerce analysis looks like.

Not the only way. Often the right one.

Pricing that scales with the number of stores you report on.

Operator

Team

Agency

Free playground + Scale

Boring where it should be.

OAuth-only

EU-hosted

Read-only by default

Questions you'll ask.

Bring your ecommerce data into one AI-ready context.