Should I use the MCP server or the REST API?
On this page
You can reach the Clinia Context Engine for Health AI through an MCP server or a REST API. Both work with the same patient data, so the choice depends on what you're building. API Overview in the developer documentation compares the two.
For an AI agent
Use the MCP server. It's designed for AI agents and gives them tools to browse, read, and search a patient's data. The data is laid out as a file tree, so an agent can find what's there without knowing its structure in advance. Connect an AI Agent walks you through the setup.
The MCP tools are tuned for agents and change with the product, so their stability is never guaranteed. An agent can also run on the REST API, but the MCP server is the one designed for it.
For everything else
Use the REST API for server-side integrations, dashboards, and sending data to the engine. Its graph endpoints also show how the engine reached a fact: Audit the Patient Graph traces a fact back to the extracted facts behind it. The graph endpoints are a technical preview, so they can change in any release.
Need help?
Not sure which one fits your use case? Write to Clinia Support and describe what you're building.