Merron

EU AI Act · Readiness and evidence

Know which questions to answer before relying on a compliance claim.

Merron helps you describe an AI product, confirm its business use and organise a practical evidence record. It makes missing facts and unsupported conclusions visible.

Start a private caseUse your AI assistant

Who is Merron for?

Founders, software teams and businesses preparing an evidence record for a specific AI product or feature. The current screen is most useful for common provider/deployer scope and transparency questions. Complex, potentially prohibited and sensitive uses remain unresolved.

What do I need to supply?

Describe the specific product, its intended purpose, who uses it, whether EU customers or outputs are involved, and your role. Add relevant product documents and user-facing disclosures. Confirm the facts you know and leave uncertain answers unknown.

What will I receive?

In the private workspace: a screening result where supported, documented observations, missing evidence, prioritised next actions and a saved PDF. Through MCP: a bounded screen of confirmed facts and observations about selected excerpts, without a saved workspace report.

Does a report prove that we comply?

No. Merron does not issue certificates, verify live controls or establish complete legal compliance. A screening result records a conditional assessment of the supplied facts. Unresolved issues must be addressed before relying on conclusions.

Can a source-code scan answer everything?

No. Code can indicate a model integration or disclosure wording. It cannot establish actual intended use, deployment, customer geography, legal roles or whether users see a notice. Those facts need confirmation and supporting evidence.

How is this different from a general AI conversation?

Merron supplies a fixed screening scope, versioned rules and an explicit record of missing facts. Its workspace adds source-linked document observations, a separate automated evidence check and saved report history. These are workflow features, not evidence that its legal accuracy exceeds another AI model.

What does the beta cost?

No payment is required. The private beta currently includes two cases, thirty copilot replies and three report generations per account, subject to daily service capacity and document limits. Public MCP tools do not consume the workspace’s AI allowance.

Read the screening scope and accuracy limits and customer guide before starting.