Yatis MCP Server

The Yatis MCP Server bridges the gap between vehicle telemetry and AI assistants by exposing a rich set of telematics APIs through the Model Context Protocol. It turns raw data from Yatis’ fleet management platform into consumable tools that can be invoked by agents such as Claude, VS Code, or Cursor. By running locally on Node 20+, the server ensures low latency and full control over credentials, making it ideal for developers who need secure, real‑time access to vehicle information without exposing keys to third‑party services.

At its core, the server offers a straightforward set of endpoints that cover every common telemetry need: listing groups tied to an API token, retrieving vehicle and device metadata, fetching current or historical location data, and pulling electric‑vehicle battery statistics like voltage and state of charge. It also provides operational metrics such as distance travelled, power consumed over time, and even the ability to calculate great‑circle distances between two coordinates using the Haversine formula. These capabilities empower AI agents to answer natural‑language queries such as “How far has vehicle #12 driven today?” or “What is the current battery level of all EVs in fleet A?” with instant, accurate data.

Developers benefit from the server’s tight integration patterns. In VS Code, a simple configuration launches the server as an stdio process, allowing extensions to invoke tools directly from the editor. The same setup works with Cursor and Claude Desktop, making it a versatile component across multiple AI workflows. Because the server communicates via MCP, agents can discover available tools automatically, construct prompts that include real‑time telemetry, and chain multiple tool calls to build complex logic—such as routing decisions based on current vehicle locations or generating maintenance alerts from battery health metrics.

Real‑world scenarios that leverage this server include fleet monitoring dashboards, autonomous vehicle status reporting, and predictive maintenance systems. For instance, a logistics company can let an AI assistant summarize daily fuel consumption across its fleet, while a delivery app could query the nearest available vehicle in real time. The server’s lightweight Node implementation and minimal dependencies mean it can be deployed on edge devices or cloud VMs without heavy overhead, giving teams flexibility in how they orchestrate AI‑driven operations.