European payment infrastructure for AI agents
Give your AI agents the ability to pay safely
Bank transfers (incl. SEPA) · Card payments · Stablecoins · via MCP, SDK, CLI or API
The problem
Your agents can already pay. Nothing governs them.
/01
No caps
They spend with the company's card or keys: no limit of their own.
/02
No identity
At payment time, nothing distinguishes a legitimate agent from a compromised one.
/03
No accountability
Which agent, which instruction, which rule? Nothing is recorded. Nobody answers for it.
When an incident occurs, nobody bears the loss today. Praesia is built to hold that role.
How it works
One mandate signed once. Agents that pay on their own.
Sign the mandate
Caps, beneficiaries, duration. One strong signature, one time.
Badge your agents
Each agent gets its own identity, tied to a human owner.
Let them pay
Every payment is checked before execution. Any overrun is refused. One click revokes everything.
Operational in a day: one transfer funds the wallet, your agents connect. No integration with your banking systems.
The product
Authority delegated, exercised, then cut off.
Root mandate
A root mandate is created and signed with SCA. Every mandate is a contract.
Delegations
Sub-mandates per team. Rights can only shrink going down.
The agent pays
Badged and connected over MCP, the agent settles an invoice. Identity, mandate, intent: checked before execution.
Sealed audit
Every decision traced: agent, prompt, rule, mandate version, owner, outcome.
Kill switch
One click cuts a branch: instant, cascading revocation. The monitoring engine can also cut on its own.
Developers
One connection. Checks before every payment.
Your agents connect over MCP, natively from Claude, GPT and agent frameworks, or through the SDK, CLI and API, with no integration with your banking systems. You configure mandates in the console. Praesia checks every payment before execution.
Guarantee
Outside the mandate = reimbursed.
Any payment executed outside the mandate engages our liability: we reimburse it. Inside the mandate, your exposure is capped by construction.
FAQ
Frequently asked questions.
Why use Praesia?
Praesia lets your AI agents pay autonomously, within a framework you control: a mandate signed once sets caps, beneficiaries and duration. Your maximum loss is known in advance, any overrun is blocked before execution, and any payment outside the mandate is reimbursed. Deployed in a day, without touching your banking systems.
Is Praesia a bank?
No. Praesia is the governance and execution infrastructure for agent payments. Funds are ring-fenced with a licensed partner.
What happens if an agent is hijacked?
Its maximum loss is bounded by its mandate, known in advance. The monitoring engine detects and suspends abnormal behaviour. If an out-of-mandate payment were executed, it would be reimbursed.
Which payment rails?
Bank transfers (incl. SEPA), per-agent cards and stablecoins for machine-to-machine. Praesia plugs into all existing payment rails.
How do my agents integrate?
Over MCP, SDK, CLI or API. The badge signs every request.
How long to get operational?
One day. Praesia does not integrate with your banking systems: you fund your wallet with a simple transfer, and your agents connect over MCP, SDK, CLI or API.
How is this different from spend-management software?
Software that does not hold the funds issues an opinion, which can be bypassed. Praesia attaches the rule to the funds themselves: it applies before every execution, and an identified counterparty answers for it.
Let your agents pay safely
Early access, founder pricing, direct influence on the roadmap.
contact@praesia.eu