InfluxDB MCP

InfluxData ships an MCP server for InfluxDB 3, so an assistant can both query and administer a time series database over the Model Context Protocol. It connects to a Core or Enterprise instance and exposes the query engine directly: an agent can run SQL against measurement tables, inspect the schema of a database before writing a query, and return rows the model can summarize, compare or chart. Administration is the other half. Tools cover the database lifecycle, including creating and deleting databases and tables, listing what exists, and creating or revoking API tokens with specific scopes. Data can also go in, with points written in line protocol, so an agent can persist values it has computed rather than only reading history. Everything runs against the instance you point it at, self-hosted or hosted, using a token you issue and can revoke. It suits platform, IoT and observability teams who store metrics or sensor readings in InfluxDB. Common uses are exploratory questions across a metric history, checking which tables and columns exist before building a dashboard query, and routine housekeeping such as provisioning a database for a new service or rotating a token that has spread too far.