Grow your ecommerce brand from one place.

Shopify, Meta, Google, TikTok, Klaviyo and GA4 together, with creative analysis built in and your AI connected. €99 a month, everything included.

14 days free. A card is needed to start, and nothing is charged until the trial ends.

Demo data: Northlight, a fictional brand.
  • Shopify
  • Meta
  • Google Ads
  • TikTok
  • Klaviyo
  • Google Analytics 4

Read-only at the source · Works with ChatGPT and Claude · €99 a month, everything included

Creative Intelligence, built in

Acquisition runs on creative, and most brands pay for a separate tool to see which ads work. In Kreel it’s part of every plan. Your Meta and TikTok ads sit in one gallery with the video or image, spend, ROAS and how each has changed, sorted by what needs attention.

Every active ad sits in Testing, Scaling, Winner, Fatiguing or Kill. Your brand’s own rules place it there, and each column prints its rule with the real numbers. Every brand starts on default rules; change any number, separately for Meta and TikTok.

The lifecycle stage board for Northlight’s Meta ads: Testing (10 ads, €5,976), Scaling (2 ads, €1,087), Winner (8 ads, €4,890), Fatiguing (6 ads, €5,490), Kill (1 ad, €827), each column printing its rule with the brand’s numbers, under the precedence line (Kill > Fatiguing > Scaling > Winner > Testing, first match wins) and the efficiency bars (Cold bar: 2.6x (target), Warm bar: 4.1x, Retargeting bar: 6.2x, Account bar: 3.5x). The lifecycle stage board for Northlight’s Meta ads: Testing (10 ads, €5,976), Scaling (2 ads, €1,087), Winner (8 ads, €4,890), Fatiguing (6 ads, €5,490), Kill (1 ad, €827), each column printing its rule with the brand’s numbers, under the precedence line (Kill > Fatiguing > Scaling > Winner > Testing, first match wins) and the efficiency bars (Cold bar: 2.6x (target), Warm bar: 4.1x, Retargeting bar: 6.2x, Account bar: 3.5x).
The lifecycle board, each column printing its rule. Demo data: Northlight, a fictional brand.

Meta and TikTok, one platform at a time. Fatigue flags and hook rate are Meta only. Revenue and ROAS are what each platform reports for the ad.

Creative Intelligence

Every channel in one place

Shopify, Meta, Google, TikTok, Klaviyo and GA4 sync into one workspace on a schedule. Open Kreel and yesterday is already there: performance against the previous period, what changed, margin after ads, customers and cohorts, products and stock, and your email campaigns and flows. There’s nothing to build. The pages and the weekly brief work as soon as your data syncs. Add monthly targets and Kreel paces the month; add product costs (Shopify’s own unit costs often cover them) and margin appears.

Performance & margin

Northlight’s Performance overview, month to date (vs Aug 1, 2026 – Aug 17, 2026): the pace line (€71.6k of €145.5k September plan · 92% of pace · 14 days left), then eight KPI tiles, net revenue €71,579 (+6.1%), orders 671 (+7.0%), AOV €106.67 (-0.9%), new customers 293 (-21.4%), ad spend €19,321 (+3.3%), MER 3.70x (+2.7%), new-customer CAC €65.94 (+31.5%), email share 23.5% (-0.5 pp), 5 of them with a plan or target chip, then the revenue and ad spend chart by day with its dashed target pace line, and the €5,280 a day it takes to reach €145,500. The month’s pace over net revenue, orders, ad spend and MER, each with its change and plan chip, for Northlight.Northlight’s Performance overview, month to date (vs Aug 1, 2026 – Aug 17, 2026): the pace line (€71.6k of €145.5k September plan · 92% of pace · 14 days left), then eight KPI tiles, net revenue €71,579 (+6.1%), orders 671 (+7.0%), AOV €106.67 (-0.9%), new customers 293 (-21.4%), ad spend €19,321 (+3.3%), MER 3.70x (+2.7%), new-customer CAC €65.94 (+31.5%), email share 23.5% (-0.5 pp), 5 of them with a plan or target chip, then the revenue and ad spend chart by day with its dashed target pace line, and the €5,280 a day it takes to reach €145,500.
Performance, paced against the month’s targets. Demo data: Northlight, a fictional brand.
The Paid ads table at campaign grain: the total, then Meta Ads with its campaigns and their audience badges. The Paid ads table at campaign grain: the total, then Meta Ads with its campaigns and their audience badges.

Paid ads

Meta, Google and TikTok in one table, from channel down to ad, with spend paced against the month’s plan.

The email a sent campaign sent, as Kreel renders it on the campaign’s page: the brand’s masthead and the email’s headline. The email a sent campaign sent, as Kreel renders it on the campaign’s page: the brand’s masthead and the email’s headline.

Email & SMS

Each email next to what it earned, the lifecycle flows you’re missing, and deliverability by inbox provider.

The median payback tile from the cohort KPIs: the month in which new customers’ revenue covered their CAC, and how many cohorts have paid back.

Customers & retention

Cohorts, CAC payback by channel, and the customers who should have reordered by now.

The critical stockout finding from the “What changed” rail: a product with fewer days of cover than the alert, the store net revenue at risk, and a button that opens Inventory. The critical stockout finding from the “What changed” rail: a product with fewer days of cover than the alert, the store net revenue at risk, and a button that opens Inventory.

Products & inventory

Days of cover and return rates by product, with the Meta spend still going to products about to sell out.

Demo data: Northlight, a fictional brand.

Your ChatGPT or Claude, briefed on your brand

Connect Kreel to ChatGPT or Claude once, then ask anything about the brand. Your AI reads the same numbers as the dashboard. Fill in Brand Knowledge, or let your AI interview you to fill it, and it also knows your targets, positioning and what you’ve learned. Ask “why did new-customer CAC jump last week?” and the answer is about your brand, with the dates it covers and how fresh each source is. See something odd on a page? Press Ask, and your AI opens with that exact view attached.

A mock Claude conversation: asked what changed last week, it calls Kreel’s get_brand_context, get_performance and get_findings tools, then answers for the last 7 days with net revenue, ad spend, MER and new-customer CAC in one table, each platform’s own reported revenue and ROAS in another, never added up, the date the data runs through, and three things to do first: replace a wearing-out Meta ad, reorder a product about to sell out, and close the month’s revenue gap. A mock Claude conversation: asked what changed last week, it calls Kreel’s get_brand_context, get_performance and get_findings tools, then answers for the last 7 days with net revenue, ad spend, MER and new-customer CAC in one table, each platform’s own reported revenue and ROAS in another, never added up, the date the data runs through, and three things to do first: replace a wearing-out Meta ad, reorder a product about to sell out, and close the month’s revenue gap.
Your own Claude, answering from Kreel. Demo data: Northlight, a fictional brand.

There’s no chat inside Kreel: the analyst is your own ChatGPT or Claude. In ChatGPT, install Kreel from the plugin directory (availability depends on your plan, workspace and region); in Claude, it’s a custom connector with one URL.

Your AI & Brand Knowledge

€99 a month, everything included

One brand, every source and every feature, after a 14-day free trial. Plans differ only in brands, seats and support, never in features, and the price never depends on your store’s revenue. One subscription instead of a dashboard tool, a creative tool and a reporting tool.

14 days free. A card is needed to start, and nothing is charged until the trial ends.

Running several brands?

Agencies and multi-brand teams get every brand under one roof, each with the full depth: Creative Intelligence, every channel and your own AI. On top, Portfolio ranks the brands by what needs attention, and client reports go out as a link with your logo and colours.

Kreel for agencies

Read-only at the source

Kreel never writes to Shopify, Meta, Google, TikTok, Klaviyo or GA4. When you connect ChatGPT or Claude, you choose read only or read and write. Write access stays inside Kreel: your AI can update Brand Knowledge, load costs or a plan and publish reports, but it can’t touch your store, ad accounts or Klaviyo. Customer names, emails and addresses never reach your AI, and every call it makes shows up in the agent activity log.

How Kreel handles your data

The Kreel consent screen ChatGPT opens when it connects, “Allow ChatGPT to access your Kreel data?”: “read your connected brands’ performance data” is required, “update your brand knowledge (profile, sections, competitors, learnings) and create drafts / upload assets” is a ticked box the person can untick, then Allow and Deny. The Kreel consent screen ChatGPT opens when it connects, “Allow ChatGPT to access your Kreel data?”: “read your connected brands’ performance data” is required, “update your brand knowledge (profile, sections, competitors, learnings) and create drafts / upload assets” is a ticked box the person can untick, then Allow and Deny.
The consent screen your AI goes through to connect. Demo data, fictional brands.
One call in the agent activity log, opened: Claude’s query on Northlight, outcome ok, run by jonas.berg@harborandco.example over MCP, with its latency and response size, then its arguments, a query of daily_channel_summary for channel, spend and revenue by channel since 2026-08-20, under “Secrets and emails are redacted and long values shortened before storage.” One call in the agent activity log, opened: Claude’s query on Northlight, outcome ok, run by jonas.berg@harborandco.example over MCP, with its latency and response size, then its arguments, a query of daily_channel_summary for channel, spend and revenue by channel since 2026-08-20, under “Secrets and emails are redacted and long values shortened before storage.”
Every call your AI makes, in the agent activity log. Demo data: Northlight, a fictional brand.

Built inside an ecommerce growth agency

We’ve managed eight figures in ad spend across Meta, Google and TikTok for growing brands. That meant a tab for every platform, dashboards built by hand and reports to send, so we built one tool to make decisions from. It worked well enough that we opened it to other brands and agencies.

Why we built Kreel

Questions about Kreel, answered plainly

All questions

Do I need to connect every platform?

No. One source is enough to start, and each page says which source it needs. Anything that isn’t connected says “not connected” instead of showing a zero.

Which ads does Creative Intelligence cover?

Meta and TikTok, one platform at a time, because the two attribute differently. Google Ads has no creative view. Ads whose video or image the platform doesn’t expose (some TikTok Spark Ads, live and catalog formats) still show their numbers and stage, with a placeholder instead of the media and no tags.

Which parts need Claude or ChatGPT?

Asking questions in your own words, the ready-made prompts, AI-written reports, the test ledger and setting things up from a chat. The dashboards, Creative Intelligence (its AI tags are made by Kreel, not by your AI), the weekly and daily briefs, publishing a brief as a report, digests, alerts, scheduled reports and approvals all work without one.

Does Kreel change anything in my ad accounts or store?

No. Kreel only reads from Shopify, Meta, Google, TikTok, Klaviyo and GA4 and never writes to them. It doesn’t create, edit, pause or send anything there.

Is there a free trial?

Yes, 14 days on any plan. A card is needed to start, and nothing is charged until the trial ends. Cancel before then and you pay nothing.

Is Kreel an attribution tool?

No. Kreel isn’t an attribution tool. It uses what each platform reports, next to what your store recorded, and keeps them separate. Most teams run the business on those numbers anyway. Kreel is built for the work that comes after.

Grow your ecommerce brand from one place.

14 days free. A card is needed to start, and nothing is charged until the trial ends.