Skip to main content
Layer’s MCP server exposes Layer accounting data and bookkeeping workflows as Model Context Protocol tools. AI clients and platform agents can use these tools to inspect financial data, answer business questions, and take supported bookkeeping actions through Layer’s API.

​
Insights

The MCP server supports two operating modes for insight generation:
  • Business-level insights: Tools operate within one configured business. This is the right fit when an end business owner or business-scoped AI assistant needs answers about that business’s transactions, reports, and chart of accounts.
  • Platform-level insights: Tools can set the active business for a session, allowing a platform-owned AI agent to move between businesses programmatically. This is useful for platform support, operations, demos, and internal agents that already know which Layer business they should inspect.
In platform mode, the agent calls set_active_business before using financial tools. In business-scoped mode, business context is fixed by the MCP URL and the active business is not exposed as a tool. Layer’s MCP server is designed for two integration patterns:
  • Platforms can incorporate Layer MCP tools into their own AI agents, giving those agents access to Layer financial data and bookkeeping actions for the platform’s end customers.
  • Platform owners can use an MCP server to draw insights about their customers, either from their own internal agent or from an MCP client such as Claude Desktop or Claude on the web.