Back to FAQ
Getting started with MCP
Metry’s MCP server lets any MCP-compatible AI tool query your energy data directly - no copy-paste, no exports, no API keys to manage.
MCP (Model Context Protocol) is an open standard that lets AI assistants connect to data sources and pull live data in real time.
What you can do
Once connected, you can ask your AI tool questions like:
- “Which building used the most electricity last month?”
- “Which meters have not reported data this week?”
- “Compare heating consumption for property X between this year and last year.”
- “Flag any meters with missing data since Monday.”
The AI pulls from your actual Metry data rather than making assumptions.
Connect your AI tool
The MCP server URL is: https://mcp.metry.io
Authentication is handled via OAuth when you first connect. Go to Integrations in the Metry portal for step-by-step setup instructions for each supported tool.
Supported tools
| Tool | Setup command |
|---|---|
| Claude (Desktop/Web) | Add directly from the Claude connectors directory |
| Claude Code CLI | claude mcp add --transport http --scope user metry https://mcp.metry.io |
| Gemini CLI | gemini mcp add --transport http --scope user metry https://mcp.metry.io |
| Codex CLI | codex mcp add metry --url https://mcp.metry.io |
| Antigravity CLI | agy mcp add metry https://mcp.metry.io |
| Mistral Vibe | Add via Context > Connectors > Add custom connector |
| ChatGPT web | Add via Plugins > Browse Plugins > add custom connector |
You need at least one organization with meters to use the MCP server.