Identity verification and fraud prevention company iDenfy has introduced a Model Context Protocol (MCP) connector that links Claude and Claude Code to a business's iDenfy account. The connector allows teams to retrieve verification details, such as KYC and the most common denial reasons, through plain-language requests in the AI assistant. Users no longer need to navigate the iDenfy dashboard for this information, and sensitive data such as biometrics and ID document photos are not passed to the AI model.
How the connector handles data
iDenfy has built the connector on dedicated endpoints that have no access to personal data. These endpoints return only statuses, dates, and other non-personal information, which means names, document numbers, and photos cannot be passed to Claude. For AML screening, answers show statuses and detected flag types rather than matched names. In addition, the connector runs on a separate access token, isolated from iDenfy's core system interfaces.
Control over what the AI assistant can see remains with the business. Customers select 'View' or 'Edit' access separately for each data area, and iDenfy checks permissions on every request. Connections can be removed from the dashboard at any time, while unused connections expire automatically after seven days.
Use cases for compliance and operations teams
According to the company, compliance teams in fintech, crypto, iGaming, and other regulated sectors can use the connector to run queries that previously required dashboard searches and spreadsheet exports. The underlying data is unchanged; what differs is the method of access. Adomas Vitkauskas, Chief Product Officer at iDenfy, noted that compliance teams need fast answers from verification data while end users' identities remain protected.
The connector also extends to other functions. Finance teams can check how many verification units remain and when they are likely to run out, while product teams can identify which document types have the lowest approval rates and where users drop off, and apply those findings to the onboarding flow.
The launch comes as businesses weigh the risks of connecting AI tools to sensitive systems. In a 2026 survey by Gallagher, 55% of businesses named data protection and privacy violations among the top perceived risks of AI adoption. At the same time, Deloitte predicts that by 2027, half of companies using generative AI will have launched agentic AI pilots or proofs of concept.
MCP has become a common method of connecting AI assistants to business systems. In identity verification, however, account-level access carries additional requirements. iDenfy states that giving AI access to account data without exposing end users' identities required a separate layer of endpoints and secure authentication.
From documentation to account data
The connector builds on an MCP server that iDenfy launched in April 2026, which brought the company's public API documentation into AI assistants to support developer integrations. That server had no access to customer accounts by design. The new connector extends MCP support to each customer's own account data.
The Claude connector is available free of charge to all iDenfy customers on every plan. The company is also working on support for other AI tools, including ChatGPT, Gemini, and Grok.
