.png&w=3840&q=100)
One payment · £0 every year after · Nobody else holds your data
Stop renting your data.
Shopify, Amazon, Meta, Google, TikTok, Amazon Ads, Klaviyo and GA4 in one dashboard you own, synced automatically. Ask AI questions and get answers from your actual numbers.
Real product screens, showing the built-in Abaxar Coffee demo dataset.
Built for desktop. Have a look now, or open it later at your desk.
Every reporting tool charges you monthly to look at numbers your own store already generated. The data is yours. The dashboard never is.
And it stacks up. Shopify, email, accounting, apps, ad tools, then a few hundred a month for analytics on top, just to understand your own business. You're renting access to your own numbers, and you never end up owning the system that makes sense of them.
A few years ago, building your own reporting layer made no sense. Too technical, too expensive, too much to maintain. AI coding changed the maths.
AI can now build and run a full reporting system for your business. It pulls every channel's numbers into one place, works out profit automatically, and answers questions in plain English from your own data.
You don't need a dev on payroll or another SaaS subscription. You need a system that's yours.
That's what Abaxar is.
A reporting dashboard you own, running on your own machine, pulling from all your channels automatically, with a built-in AI data layer so you can ask questions and get answers using your actual business numbers.
Shopify and Amazon sit side by side in one place you own, not split across two logins. Amazon is a black box. Abaxar gets your seller and PPC data out of it and into a system you control.
You run your own copy using an AI coding tool like Claude Code, Cursor or Codex, and it walks you through the whole setup. Everything runs on your machine, and there's no subscription to anyone.
What the dashboard covers
Abaxar ships with tabs built around how ecommerce operators work.
Overview
Your daily command centre. Revenue, ad spend, MER, orders, AOV, and net profit (including COGS and overhead you've set) at a glance. Month-to-date pacing against your targets. Period snapshot showing today, yesterday, last 7 days, last month. Top products, top creatives, email snapshot.
Paid Media: All channels + Meta, Google, TikTok, Amazon Ads
Cross-channel spend, attributed revenue, ROAS, and CPA, with Amazon Ads (PPC) sitting alongside Meta, Google, and TikTok. Campaign-level drilldown per platform. Creative performance tab with spend, purchases, and revenue by creative. Budget vs actual pacing per campaign. Country-level breakdowns for Meta and Google.
Shopify
Profit tab (gross profit, net profit after ads, net profit after overhead). Products table. Inventory risk view. Purchase funnel down to SKU level (sessions → adds to cart → checkout → purchase).
Amazon
Your Amazon Seller data pulled out of the black box and sat next to Shopify. Orders, units, and ASIN-level sales. Returns and FBA reporting. Financial transactions for refunds, fees, and payouts, so your Amazon profit shows up in the same place as everything else.
Email (Klaviyo)
Sends, opens, clicks, revenue attributed to email. Campaign and flow breakdown.
Settings
Connect and manage your data sources. Set COGS, overhead, targets, and budgets. Configure KPI strip defaults. Manage sync state.
Ask Abaxar
Abaxar has an AI data layer built in. Connect Claude, ChatGPT, or any AI client that supports MCP, and ask about your business in normal language. The AI queries your data and answers with real numbers.
How it works: Abaxar runs an MCP server on your own machine. Your AI client queries your data warehouse directly without sending raw data anywhere else.
Put it on a schedule: the same AI tools can run these queries on a schedule, so a morning brief or a daily stock check lands without you asking. The scheduling runs in your own AI client. Abaxar is the data layer it reads from.
Bring your own AI: Abaxar isn't locked to one assistant. It's an open MCP endpoint over your own data, built on standard protocols, so the AI client you already pay for and trust connects to the same numbers. Claude, ChatGPT, Codex, and Hermes all work today, and any new MCP-compatible agent works the day it ships. There's no bundled chatbot and nothing tying you to one vendor.
Powered by MCP (Model Context Protocol), an open standard. Works with Claude, ChatGPT, Codex, Hermes, and any MCP-compatible AI client.
Extend it with AI
Abaxar ships ready to use, and it's built to be built on, so it ends up as reporting software bespoke to your business. There's no SaaS roadmap or feature request queue between you and a change you want. You own it and you control where it goes.
Add any data source
The eight connectors Abaxar ships with (Shopify, Amazon, Meta, Google, TikTok, Amazon Ads, Klaviyo, GA4) are the starting point, not the limit. Because Abaxar is running on your own infrastructure with a consistent connector pattern, adding a new data source means asking your AI coding agent to build it. If a platform has an API, you can bring it in. eBay, Walmart, Bing Ads, Recharge, Gorgias, accounting and fulfilment systems, supplier lead times, a custom warehouse table, a Google Sheet, whatever your business runs on.
“Add a Recharge connector and pull subscription revenue into the Overview tab”. The agent reads the existing connector pattern and builds it.
Make the UI completely yours
The dashboard is built on a consistent component system (shadcn/ui), which means your AI coding agent understands the structure and can make changes cleanly. Reskin it in your brand colours. Rename tabs. Rearrange the layout. Add a new tab for a metric set that matters to your business. Surface any of the 98 pre-defined metrics already in the system, or define your own KPIs with your own formulas. Nothing has to say “Abaxar” or look like anything other than your own internal tool.
“Rebuild the Overview tab to lead with gross margin instead of revenue, in our brand colours”. Done without touching the data layer.
Extend the AI data layer
The MCP server ships with 16 tools covering the core reporting surface. But it's yours to extend. Add a new tool and your AI assistant gains a new capability: a deeper product analysis, a custom cohort query, a bespoke P&L view. The more tools you add, the more your AI client understands about your specific business.
Add connectors · Customise the UI · Extend the MCP
How it works
Step 1
Get Abaxar
After paying you connect your GitHub account, which takes a couple of minutes, and you are added to a private repository as a read-only collaborator. Clone it and follow the guided setup. Up and running in about two hours.
Step 2
Connect your data (one-time, guided)
The setup guide walks you through all eight platform connections, one at a time. Amazon, TikTok and Google Ads each carry a provider verification step that takes several days. Your API keys go into your own copy and nobody else holds them, because Abaxar connects to each platform directly rather than through a middleman. After setup, Abaxar syncs your data automatically while the dashboard is running. Hourly by default, and you can change the interval. No manual runs.
Example first session:
0–15 min: Clone your copy and open it in your AI coding tool
15–45 min: Connect Shopify and Amazon, run a test sync
45–90 min: Connect Meta + Google + TikTok + Amazon Ads
90–120 min: Connect Klaviyo + GA4, verify key metrics populate
One-time tradeoff: more setup than a SaaS signup, but after that no monthly bill and you own the system.
Step 3
Use it daily. Extend it when you want.
Open the dashboard, check the snapshot, act. When you want to change something, ask your AI coding agent. New KPIs, new connectors, layout changes. Each one takes minutes, not months.
I'm not selling you another SaaS subscription. I'm selling you the power to run your own reporting infrastructure.
You control the APIs. You control where the data lives. You're not locked into someone else's roadmap, someone else's pricing, or someone else's decisions about what data you can access.
- • No monthly subscription. One payment, use it indefinitely.
- • Not locked into a vendor. Your data lives in your own database.
- • Not limited by a SaaS feature set. Extend it however you want.
- • Not tied to one AI vendor. Bring your own AI. Point Claude, ChatGPT, Codex, Hermes, or any MCP-compatible agent at your data.
Rich Vaughan · Twelve years in ecommerce
What this costs everywhere else
€2,721.60. One year of Supermetrics, 2022. I was running my own ecommerce business and managing client ad accounts at the same time, and it was the only way to stay on top of the reporting without it taking over the day. It moved the data into a sheet and stopped there. The spreadsheets, the formulas and the reading of it were all still mine.
After twelve years in ecom I have used a lot of these tools and paid for plenty of them. Supermetrics, Triple Whale, Sellerboard and Lifetimely among them. I built Abaxar because it's the tool I've always wanted.
| Tool | What it does | Year one | Every year after |
|---|---|---|---|
| The platforms | |||
| Full dashboards with an AI layer. They price against your revenue, your ad spend or how many channels you connect, so the bill grows as you do. | |||
| Northbeam | Attribution and creative analytics | $18,000Starter, from $1,500 a monthnorthbeam.io/pricing | $18,000 |
| Triple Whale | Dashboard, AI operator, automations | $7,490Automate, annual prepaidtriplewhale.com/pricing | $7,490 |
| Peel Insights | Retention and cohort analytics | $5,388Essentials, $449 a month billed annuallypeelinsights.com/pricing | $5,388 |
| MerchantSpring | Marketplace analytics and dashboards, 120+ channels | $3,828Seller, $319 a month billed annuallymerchantspring.io/pricing | $3,828 |
| Triple Whale | Dashboard, attribution, custom metrics and SQL | $2,190Foundation, annual prepaidtriplewhale.com/pricing | $2,190 |
| Lifetimely | Profit, LTV and retention analytics, Shopify-first | $1,788M, $149 a month, for 501 to 3,000 orders a monthuseamp.com/pricing | $1,788 |
| Polar Analytics | Dashboard, attribution, AI agents, 45+ connectors | No published pricebook a demo, quoted against your GMVpolaranalytics.com/pricing | No published price |
| The pipes | |||
| They move your data somewhere you can use it and stop there. You still build the thing that reads it, and you still pay for wherever it lands. | |||
| Funnel.io | Marketing data pipeline, priced by connectors and refresh rate | $3,600Starter, from $300 a month billed annuallyfunnel.io/pricing | $3,600 |
| Supermetrics | Seven sources into a sheet or a warehouse | €1,908Growth, €159 a month billed yearlysupermetrics.com/pricing | €1,908 |
| Improvado | Data pipeline with an MCP tier for AI clients | $1,200MCP Only, $100 a month. Advanced and Enterprise are quotedimprovado.io/pricing | $1,200 |
| Windsor.ai | Seven sources, unlimited destinations | $1,188Standard, $99 a month billed annuallywindsor.ai/pricing | $1,188 |
| The single-job tools | |||
| Cheap, and good at the one thing they do. None of them puts your whole business in one place, which is the job this page is about. | |||
| TrueProfit | Shopify profit only | $420Basic, $35 a monthtrueprofit.io/pricing | $420 |
| Sellerboard | Amazon profit only | $179Standard, billed annuallysellerboard.com | $179 |
| Abaxar | Every channel in one place, on your own machine, with the AI layer built in | £249one payment | £0and every year after that |
Nearly all of them have a free or entry tier. These are what the real product costs once your business is actually running on it, and every figure is the cheapest published route to a year, which almost always means paying for the year up front. Monthly costs more. Read off each vendor's own pricing page on 2026-07-25, with the page listed under every number so you can check it. Dollars, euros and pounds as each of them publishes, not converted, because an exchange rate I picked is not a number worth arguing about. Where a company does not publish a price, it says so rather than guessing.
What they charge extra for
The prices above are the base plan. This is the part you find later, and it is where the gap gets wide, because the things sold as upgrades are the things that follow from simply owning the software.
| Capability | Elsewhere | Their price | Abaxar |
|---|---|---|---|
| Your data in a warehouse you own | Triple Whale, Data Warehouse Sync add-on, at $0 to $250K GMV | $4,990 / year | It is the product. Your database is a file on your machine |
| Custom metrics, your own SQL | Triple Whale, Foundation tier and up | $2,190 / year | Your database and your code. Define anything you like |
| Amazon | Lifetimely charge for it on top of any paid plan, for sales data | +$75 / month | Amazon Seller and Amazon Ads, both included, in the same database |
| Every extra data source | Supermetrics and Windsor.ai both tier by how many you connect | A higher plan each time | Every connector, present and future, included |
| The AI layer, once it is metered | Lifetimely say AI usage above a monthly allowance "will be billed by tokens" in the near future | Included today | Runs on your machine, against an AI subscription you already pay for |
| What happens when your revenue doubles | Priced against GMV, ad spend or monthly orders, so the bill follows you up | Your bill goes up | Nothing happens. It is already yours |
Several of these connect to Claude and ChatGPT now, the same way this does. The difference is who holds it. Theirs sits on their servers and answers for as long as you keep paying. Nobody holds yours but you.
Which is the question worth asking about any of them. If you stop paying, what do you keep? With everything above, the dashboard goes and your history goes with it. Abaxar is a copy of the software sitting on a machine you control. There is nothing to stop paying.
Get Abaxar
£249
£0 every year after
One payment. £249 for the first 50 buyers, then £349.
| Connectors | Every data connector, present and future, plus all updates, owned forever |
|---|---|
| Amazon | Amazon Seller and Amazon Ads, included |
| AI layer | MCP server on your machine, works with any MCP-capable client |
| Licence | Covers your company and the brands under the same ownership |
| Delivery | Read-only access to a private GitHub repository, granted once you connect your GitHub account after paying. A free GitHub account is required |
| Setup | A step-by-step video for every connector. Message me if you get stuck |
| Subscription | None. There is nothing to cancel |
| Refund | 60 days, no questions |
Not ready to buy? Try the live demo first.
All 50 founding places are open. The rise to £349 is announced in advance rather than sprung on anyone.
Or have it built for your whole team.
Deployed to your own cloud, every platform connected or walked through, your team's AI wired in through the MCP, and one dashboard the whole team works from. An Abaxar licence is part of the fee, so you own the software either way.
FAQ
Is the setup realistic for non-technical people?
The guided setup is designed for someone comfortable following a step-by-step guide, not a developer. You'll run the setup inside an AI coding tool (Claude Code, Cursor, Codex), and the agent does most of the heavy lifting. Most of the two hours is connecting accounts on platforms like Meta and Google, which takes time regardless of tool.
Will it run on my computer?
Mac, Windows, and Linux. Abaxar is one codebase that runs locally on all three, so the question is less which machine you own and more what is on it: Node 20.9 or newer, Git, and an AI coding tool like Claude Code, Cursor or Codex. If yours turns out to be the exception, that is what the refund is for.
What do I actually need to connect?
The accounts you want to pull data from: Shopify, Amazon, Meta Ads, Google Ads, TikTok Ads, Amazon Ads, Klaviyo, and GA4. You're guided through each one. Abaxar ships with those eight. If a platform has an API, you can add more connectors over time.
Where does my data live?
Nobody else holds it. Abaxar keeps everything in a single database file on your own computer, and your API keys sit on that same machine. There is no Abaxar server in the path, so Abaxar never stores or accesses your data. You control it.
What if I get stuck?
Your AI coding agent can read the codebase and walk you through any step, and in most cases it can just do it for you. The setup guide covers every connector, and there are video walkthroughs.
Does it really get my Amazon data out?
Yes. Abaxar connects to the Amazon Selling Partner API and the Amazon Ads API and pulls your orders, units, ASIN-level sales, returns, fees, and PPC performance into the same database as your Shopify data. Amazon normally keeps this locked in Seller Central. Getting it out into a system you own, sitting next to Shopify, is one of the main reasons operators buy Abaxar.
Can I customise it?
Yes. This is the main advantage over a SaaS. Ask your AI coding agent to add a new KPI, connect a new data source, reskin it in your brand, rebuild a section, or build something entirely new. Many changes are live in minutes.
What about platform API changes?
Connectors are built to stay maintainable. When a platform changes an API, you update the connector by asking your AI coding agent. Lifetime updates are included.
Why not just use a SaaS analytics tool?
You might already be using one. If you're happy with it, keep using it. Abaxar is for operators who want to stop paying monthly for something they can own, or who want to go further than what a SaaS will let them do. The setup tradeoff is real. A SaaS is faster to start. But after that one-time setup, there's no monthly bill and no ceiling on what you can build.
What does it cost to run?
Nothing to run Abaxar itself. It runs on your own computer and keeps your data there, so there is no hosting bill and no database bill, whatever the size of your catalogue or however many years of history you keep. You bring your own AI client, which you are already paying for. There's no Abaxar subscription either.
Refund policy?
Full refund for 60 days, no questions.
Specs for nerds
The rest of this page tells you what Abaxar does. This is the spec sheet, for the reader who wants to see the engine before buying.
Stack
- Framework: Next.js 16, App Router
- Language: TypeScript end to end
- UI: shadcn/ui on Tailwind CSS 4, Recharts for charts
- Runtime: Node 20.9 or newer
- Platforms: macOS, Windows, Linux
- Setup: clone, install, migrate, run
Database
- Engine: one libSQL file on your machine, file:./local.db
- Schema: built from empty by a chain of SQL migrations
- Pipeline: raw platform data → fact tables → display contracts the dashboard reads
- Infrastructure: no database server, no provider account
Sync
- Scheduler: in-process, inside the running dashboard
- Jobs: each sync spawned as a detached child process
- Interval: off / hourly / 3h / daily, hourly by default
- On demand: the Catch up now button
- Window: syncs while the dashboard is running
AI data layer
- MCP server: built in, at /api/mcp
- Tools: 19 read tools: 16 core, 3 from connector modules
- Clients: anything that speaks MCP
- Extension: add your own tools through the same module seam
Correctness
- Metrics: a registry defines every KPI once
- Anti-drift: npm run metrics:verify keeps the formula shown equal to the number computed
- Tests: a Vitest suite and CI on every change
Delivery and updates
- Access: a private distribution repository, you as a read-only collaborator
- Updates: git pull, with the running version and an update check in the dashboard
- Your data: local.db and .env.local are gitignored, so updates never touch them
Security posture
- Credentials: plaintext in .env.local, behind your operating system's file permissions
- At rest: not encrypted
- Custody: keys go from your machine to Shopify, Meta, Google, TikTok, Klaviyo, GA4 and Amazon, and nowhere else
- Third parties: no server of ours in the path
- Network: the app listens on 127.0.0.1
- MCP auth: none, by design; a request guard rejects non-local calls
- Boundary: your machine is the security boundary
- Beyond localhost: put real authentication in front of it first. The README says the same in the same words
Get Abaxar.
Your numbers in one place, in a system you own, with every connector and the AI data layer already built in. There's no subscription behind any of it. It's the first piece of an owned ecommerce stack.
£249 one-time · No subscription · Full refund for 60 days, no questions · Lifetime updates included