Mercur

Run a marketplace without taxing your own growth

Mirakl charges a percentage of everything you sell, for as long as you sell it. Mercur runs on the opposite model – own the code and the data, host it on your own infrastructure, and keep what you build.

Open sourceYour infrastructureZero GMV fees

Same problem, opposite ends

Mirakl and Mercur both put third-party vendors on your store – from opposite ends. Mirakl is a closed SaaS you rent. Mercur is an open platform you own. Here is what owning it gives you.

Zero GMV fees

The license is your only platform cost. Your success isn't taxed.

Own the source

Open source, full source for licensees – an engine you can read, audit, and extend.

Your infrastructure

Any cloud, on-premise, or air-gapped. Data residency is your decision.

Built on the most popular framework

Standard JavaScript and TypeScript – any web developer can work on it, no certified specialists to hire.

Additive, not disruptive

Add a third-party layer without replatforming, or stand up a full marketplace from scratch.

No vendor lock-in

Keep the version you paid for and everything built on it – forever.

Mirakl vs Mercur, by architecture

The differences that decide this are structural, not feature-by-feature.

MiraklMercur
Core technologyClosed, vendor-onlyOpen source (MIT), public
Enterprise codeBlack boxFull source for licensees
IntegrationsSeparate productsBuilt into one platform
CustomizationTickets and workaroundsYour own layer, official extension points
DeploymentVendor cloud onlyAny cloud or on-premise
Cost modelA percentage of your GMVZero GMV fees
ExitLock-inYou keep what you paid for
Developer poolCertified specialistsAny web developer

Your success shouldn't raise your bill

A percentage of GMV means the better your marketplace does, the more you pay – for the same platform. Break-even gets measured in years, and the meter runs before your first third-party sale. Mercur charges for deployment and support. Your growth stays yours.

Percentage of GMV / MiraklFlat platform cost / Mercur
Tax
LaunchYour turnoverScale

At scale, the meter is the largest line on the bill.

We know Mirakl from the inside

We didn't reverse-engineer the market leader. We build with admins who ran Mirakl for years, across some of the largest marketplaces in the region. Here is what running it actually looks like.

1

No native BuyBox

Every large admin builds the buy-box logic themselves, outside the platform.

2

Exports leave no trace

A departing employee can pull the entire vendor base, and nothing logs it.

3

Settlement on the current period

Corrections land in the current cycle, which collides with per-period accounting.

4

Shipping you can't verify

With no carrier tracking, orders auto-mark as received after a fixed window – whatever actually happened.

When Mirakl is the right call

We don't think Mercur is for everyone.

Choose Mirakl

You don't have your own vendors and you need a ready network of 100k+ merchants live in about 90 days – and a percentage of GMV on a vendor cloud works for you.

Choose Mercur

You bring your own vendorsYou want to own the code and the dataYou need on-premise or EU data residencyYou want a cost that doesn't grow as a tax on your success

Whose data feeds your AI?

A marketplace leans on AI in more places every year – ranking search, recommending products, enriching vendor catalogs, and now answering shoppers' own AI agents. All of it runs on one thing: your catalog and transaction data. The question is where that data goes for the AI to use it.

With Mirakl

Your data → Mirakl's cloud → Mirakl's AI

The AI runs in Mirakl's cloud, on the data you route through it. The model that learns from your marketplace isn't one you host or control.

With Mercur

Your data stays → your AI, your infrastructure

Search ranking and recommendations run on your catalog, in your environment.llms.txt and an MCP server let AI agents read your store – access you control.No customer or vendor data is routed to a vendor's model to make it work.

Three ways to go deeper

Read the comparison, then follow the thread that matters to you.

See it on your own terms

Read the comparison, run the stack, then talk to us about putting it in production – on your infrastructure, your data, your terms.