AI has compressed one of the hardest parts of starting a software company.
An idea can become code, a working interface and even a deployable product in days. But the moment someone wants to pay for it, speed disappears. A founder suddenly needs checkout, subscription logic, usage metering, taxes, invoices, chargebacks, affiliate payouts and a way to make sure a heavy AI user does not consume more compute than they paid for.
Creem is betting that this second half of building a company is now the bottleneck.
The Tallinn-based startup has raised €5 million in seed funding, bringing its total capital raised to €7 million. The round was led by Inovo VC, with Practica Capital and Antler among its backers. Creem was led by Gabriel Ferraz, Toms Bergans-Berģis, Koliyanov Vyacheslav and Pavel Ruban.
The funding arrives alongside Creem 2.0, which pushes the company beyond simply processing a software purchase.
AI products often do not behave like traditional SaaS. One customer may generate ten images while another generates ten thousand. An agent may consume tokens, API calls, compute time or credits rather than occupy a conventional software “seat.” A flat subscription can therefore separate what the customer pays from what the product actually costs to run.
Creem 2.0 introduces custom meters, usage-based billing and prepaid credit wallets designed around that problem. A customer can fund credits before the compute is consumed, while developers can meter generations, tokens, API requests, storage or other units.
Creem describes the philosophy simply:
“If you built the product in a weekend, charging for it should feel like a weekend too.“
The more unusual part is who can run that infrastructure.
Creem exposes its store through APIs, a CLI and MCP, allowing an AI coding agent to create products, open checkouts, configure webhooks and monitor billing without relying entirely on a human clicking through a dashboard. The company is effectively trying to make finance infrastructure machine-operable, rather than merely adding an AI assistant on top of an existing payments product.
There is an early example of what that can look like. In a Creem-published customer case study, Uneed founder Thomas Sanlis said Claude Code used Creem’s SKILL.md integration guide to rewrite more than 1,500 lines of payment code using three prompts, including endpoints, webhooks and checkout flows. Creem reports no integration errors in that migration; because this is a company-published customer case, the result should be treated as customer-reported rather than independent benchmarking.
The same platform also handles the less futuristic work founders cannot ignore. As Merchant of Record, Creem becomes the legal seller, taking responsibility for payment processing, applicable indirect taxes, invoicing, fraud and chargebacks. Its standard public price is currently 3.9% plus $0.40 per successful transaction, with no monthly fee.
Creem says more than 1,000 AI applications already use its platform. Its next challenge is larger than processing their transactions.
AI has become capable of building the product. Creem wants the same agent to help operate the business that begins when somebody clicks “Buy.”
