MCP Zentria

A Model Context Protocol server exposing read-only Zentria e-commerce monitoring data: technical scans, alerts, intervenants, modules, sensors, sensor data history, and Temporal metrics.

The web version of Zentria is available at zentria.profileo.com.

Tools, their schemas, and usage instructions are discovered automatically by the MCP client during connection.

Endpoint

The Streamable HTTP MCP endpoint is served on POST /mcp.

Every request must contain Authorization: Bearer <ZENTRIA_MCP_TOKEN>. The public health check is available on GET /health.

Tools

Tool	Description
`get_subscription_ID_from_domain`	Finds a shop `subscription_ID` from a domain or URL. Use it first when a shop is identified by domain.
`get_subscription_ID_from_vm_hosting_name`	Finds Zentria subscriptions associated with a VM name. Use it first when a shop is identified by VM.
`get_scan_full_latest`	Returns the latest complete scan: current technical state, score, and failures. Use it for a quick diagnosis.
`get_scans_history_light`	Returns lightweight scan summaries for a period to analyze an evolution, break, or degradation.
`get_scan_full_by_ID_scan`	Returns complete detail for a historical scan identified by `scan_ID`, usually returned by `get_scans_history_light`.
`get_notifications_history`	Returns alerts and notifications emitted during a period. No notification alone does not prove the absence of a technical problem.
`get_intervenants_list`	Returns intervenants associated with a subscription.
`get_modules_list`	Returns module inventory, versions, activity, installation state, details, and update dates for a subscription.
`get_sensors_list`	Searches available sensors by keyword and returns the `sensor_ID` required for history.
`get_sensor_data_history`	Returns data history for a specific sensor. First resolve `subscription_ID`, then find `sensor_ID` with `get_sensors_list`.
`get_temporal_data_list`	Lists every Temporal metric name with its business definition, aggregation caveats, and structured-value format. Use it before history when the metric is unknown.
`get_temporal_data_history`	Returns Temporal business metrics for one or more customers. This data is independent from scans and sensors.

> Detailled parameters are detailled in the skill for AI usage

Usage Examples

These example prompts show how to ask an AI using the Zentria MCP.

“Give me the current status of shop-example.com”
“Show me the LCP trend for vmExemple over one week”
“Compare the revenue of shop-example.com for the last two consecutive months”

> To see how the AI will provide you with the results, you can consult the full process within the AI skill.

Important usage rules

Never invent a subscription_ID, scan_ID, or sensor_ID.
Dates use YYYY-MM-DD; one call is limited to 90 days.
Respect pagination: a first page may not contain all available data.
Scans describe technical state; notifications describe emitted alerts.
Temporal data is business metrics: a business variation does not prove a technical incident.
Distinguish observed facts, hypotheses, and what remains to be verified.

Client configuration example

Syntax depends on the MCP client. This is a generic example for a client supporting Streamable HTTP and headers:

{
  "mcpServers": {
    "zentria": {
      "url": "https://zentria-mcp.profileo.com/mcp",
      "headers": {
        "Authorization": "Bearer <ZENTRIA_MCP_TOKEN>"
      }
    }
  }
}

Never place a real token in a repository, conversation, or public URL.

Agent skill

The MCP is self-contained: its instructions and tool descriptions are enough to use it. The skill is an optional complement for agents that can load skills and need a more detailed Zentria workflow.

Download SKILL.md

The skill contains no secret and does not replace MCP server configuration or its access token.