Masttro, the New York wealth reporting platform used by more than 400 family offices, advisers and financial institutions, launched the Masttro MCP Server and a Developer Portal on 16 November 2026, letting clients with API access connect their data to external AI tools at no extra charge. Fintechbits checked Masttro’s own MCP page, which lists five data categories (Clients, GWM Structure, Transactions, Holdings and Performance) and says the server has been tested with Claude and ChatGPT. The connector is read-only and switched off until a client turns it on.

Availability: From 16 November 2026 for clients whose agreement includes API access

Price: No additional Masttro fee at launch

What the Masttro MCP Server Exposes

Model Context Protocol is an open standard that lets an AI application request data from another system through a defined set of tools. Masttro has built one on top of its existing API, so the AI tool does not need a custom integration for each client. A client’s approved app can read holdings, transactions, performance and the ownership and entity structures that sit behind them, including trusts and beneficiaries.

That last item is the real prize. Plenty of platforms can show a position. Few can show which entity owns it, through which trust, for whose benefit. Masttro has built its business on that mapping of complex wealth, and an AI model that can read the structure, not just the balance, can answer questions a generic portfolio feed cannot.

Masttro’s page also says setup needs no developer. A Super Admin enables access for selected people, and each user signs in with their Masttro credentials. Every request runs under that user’s existing permissions, so an analyst who cannot see a trust in Masttro should not see it through an AI tool either.

How Access Is Controlled

The release describes access as disabled by default and activated only by explicit client authorisation. Administrators can revoke it at any time. The server cannot modify records or start transactions.

Co-founder Domingo Viesca framed the product around governance. “Masttro provides that foundation for complex wealth,” he said, after arguing that the value of AI lies in the quality and context of the information a model can reach. He added that “Our MCP Server gives clients the flexibility to authorize external AI tools when appropriate.”

Fintechbits reads that restraint as the right call. Family offices are private by temperament, and a connector that defaults to off and cannot write is one a chief operating officer can defend to a principal. The firm also holds ISO/IEC 27001:2022 certification, according to its data security page, which says the certificate is available on request.

The limit is stated plainly on Masttro’s own page. Once data reaches an external AI tool, it leaves Masttro’s security perimeter, and the AI provider’s retention and training terms apply. Masttro’s access controls end where the model provider’s begin.

Developer Portal and the Masttro Intelligence Split

The Developer Portal holds the documentation, authentication guidance and integration examples. Its REST API is read-only, with 15 endpoints and short-lived bearer tokens that last 60 minutes, exchanged for an API key and secret. Identity teams can register Masttro with Microsoft Entra ID, Okta or another OpenID Connect provider.

Masttro also separates the connector from its own AI product. Masttro Intelligence is the model-agnostic assistant that runs inside Masttro’s governance framework. The MCP route is the optional exit for clients who want to use an outside tool instead. Masttro is hedging: it keeps the governed option in house and lets clients take data out if they insist.

Why the Masttro MCP Server Matters

Masttro is a specialist, and the launch will not move markets. It still matters because it shows where a niche wealth platform expects AI demand to go. Advisers who want a chatbot to read a portfolio will find a way, and a sanctioned, permissioned and revocable route is safer than an export pasted into a prompt.

The weak spot is scope. Read-only access means an AI can summarise and compare, but not reconcile, correct or act. For the accounting use case Masttro pursued with Truewind, which Fintechbits covered in its September partnership report, Fintechbits reads that workflow as running through the dedicated integration, not through a read-only MCP connection. Brokerages are making similar moves, as in the Schwab MCP server pilot, but a custodian and an aggregator face different trust questions.

The test is whether clients switch the connector on. A default of off protects Masttro, but it also means adoption depends on each family office accepting the perimeter warning above.

Frequently asked questions

What is the Masttro MCP Server?

The Masttro MCP Server is a read-only connector that lets approved AI applications retrieve a client’s Masttro data through the Model Context Protocol. It launched on 16 November 2026 alongside a Developer Portal.

Does the Masttro MCP Server cost extra?

Masttro says there is no additional charge at launch for clients whose agreement includes API access. Usage costs in the external AI tool are paid to that AI provider.

Can an AI tool change data in Masttro?

No. Masttro describes the server as strictly read-only, so connected tools cannot edit records or initiate transactions. Administrators can also revoke access at any time.

What to Watch Next

Three things will show whether the Masttro MCP Server becomes a working tool. First, whether Masttro adds write-back or reconciliation tools after the read-only start. Second, whether it publishes API rate limits, which the Developer Portal pages Fintechbits read did not state. Third, whether client administrators publicly describe how many of their users enable the connector.