Connect Yardi Data to AI and Claude with Leni

There’s no magic button for connecting Yardi data to AI. The right method depends on the customer’s authorization, Yardi setup, approved extraction method, refresh cadence, and property scope. Once those pieces are in place, Leni can normalize the agreed data through a customer-authorized ETL process and make approved capabilities available through its API or remote MCP runtime.

When the Leni path is useful

Yardi now offers Virtuoso, including its own connectors for bringing permission-aware Yardi context to supported AI tools such as Claude. The Leni path addresses a different architecture: an organization wants to use an approved Yardi ETL alongside other authorized systems, a cross-system Universal Data Model, governed organizational context, and Leni’s runtime for custom applications or compatible assistants.

A customer might choose this approach when it needs to:

• Combine supported Yardi data with other property, finance, market, document, or internal data

• Apply one governed set of definitions across systems

• Build a custom real-estate AI product through an HTTP API

• Use Leni context and analysis capabilities inside a compatible MCP host

• Keep model choice separate from the data and workflow layer

• Run durable reporting or analysis workflows beyond a single prompt

The appropriate path depends on the customer’s Yardi environment, rights, product configuration, and business requirements.

How the connection works

1. Confirm customer authorization

The customer and implementation team identify the Yardi data the customer is entitled to use, the permitted extraction method, and the users and properties that should be in scope.

2. Establish the ETL path

The approved data is extracted and transferred to Leni using the agreed ETL mechanism. This is not presented as real-time or bidirectional unless the implementation specifically supports and approves that behavior.

3. Normalize supported records

Leni maps supported source data into the Universal Data Model and applies governed semantic definitions. The exact objects and history depend on the implementation.

4. Add the customer’s context

Relevant organizational definitions, reporting practices, documents, and user context can be consulted when plan and permissions allow. Durable memory changes remain explicit user actions.

5. Deliver the workflow

• A server-to-server application can use Leni’s HTTP API with a project API key.

• A compatible assistant can use Leni’s remote MCP endpoint with Leni OAuth.

• Claude users should follow the current connector instructions in the Leni developer documentation.

What teams can do with approved Yardi data in Leni

Depending on data scope and configuration, a team can build workflows for:

• Portfolio and property performance review

• Budget-versus-actual analysis

• Month-end reporting support

• Cross-system portfolio questions

• Market and operating analysis

• Draft narratives and memos for human review

• Custom AI applications that need real-estate semantics and organizational context

These examples are not a promise that every Yardi module, field, report, or workflow is supported.

Connect Yardi-derived data to Claude through Leni

After the approved Yardi data is available in Leni, a user can connect Claude to Leni through the supported remote MCP workflow:

1. Add the Leni remote connector using https://mcp.leni.co/mcp.

2. Sign in with Leni OAuth.

3. Review the active organization and requested permissions.

4. Use the permitted Leni tools within the authorized organization and property scope.

This connects Claude to Leni—not directly to the customer’s Yardi environment. Leni supplies only the data and capabilities the customer has authorized and the user is entitled to access.

Security and control

• Source access begins with customer authorization.

• Leni enforces organization and property scope.

• Direct semantic-layer portfolio tools exposed through MCP are read-only.

• Clients do not submit raw SQL through the public MCP tools.

• The customer’s original systems remain the source of truth.

• Project API keys are for server-to-server applications, not AI assistants.

Frequently asked questions

How can Leni connect my Yardi data to AI?

We use a customer-authorized ETL process to bring the agreed Yardi data into Leni. From there, we can normalize it with your other approved sources and make the resulting data and workflows available through our API or remote MCP. Before we begin, we confirm your data rights, property scope, extraction method, history, and refresh needs with you.

How does the Leni approach compare with Yardi Virtuoso?

Virtuoso is Yardi’s own AI experience. Leni is designed for teams that want Yardi-derived data to work alongside other approved systems, documents, market data, and their own business definitions in one cross-system AI layer. We’ll help you evaluate which architecture fits the workflow you are trying to build.

Can I use Yardi data with Claude through Leni?

Yes, once the approved data is available in Leni. A compatible Claude environment can connect to Leni through our remote MCP endpoint and Leni OAuth. Claude connects to Leni and can use only the tools and data available to that user and organization.

Which Yardi data can Leni work with?

The answer depends on your Yardi setup and the data you are authorized to use. During implementation, we confirm the exact records, properties, history, transformations, and reporting cadence with your team instead of promising blanket support for every Yardi module.

Will Leni change anything in Yardi?

Not through the public portfolio tools described here. Those tools are read-only, so Yardi remains your system of record. If you need a write-back workflow, we’ll scope it separately with the required technical, security, commercial, and legal review.

Discuss the right architecture for your Yardi data

Talk to Leni about your approved Yardi data path, other source systems, target AI experience, and permission requirements. Developers can also review app.leni.co/developers/docs.