InfluxDB 3 MCP Server

Content

The InfluxDB 3 MCP Server is the official Model Context Protocol server maintained by InfluxData. It connects LLMs and agentic clients directly to an InfluxDB 3 instance — Core, Enterprise, Cloud Serverless, Cloud Dedicated, or Clustered — to run SQL queries, write line-protocol data, inspect databases and measurement schemas, manage databases and tokens, and check cluster health.

The project is written in TypeScript, distributed as the npm package @influxdata/influxdb3-mcp-server (v1.3.0), and can also run as a Docker image built from the repo. It runs over stdio by default and is configured entirely through environment variables, so it drops into Claude Desktop, Cursor, VS Code, and other MCP clients with a small config block. Full documentation lives in the InfluxDB 3 admin docs.

Source Review

The following real repository and package sources were fetched and reviewed for this entry:

• README.md — prerequisites, the full tool/resource/prompt list, per-product environment variables, and the stdio/npx/Docker client config blocks.
• npm: @influxdata/influxdb3-mcp-server — package name and version (1.3.0), license (Apache-2.0 OR MIT), engines.node >=20.11.0, and the influxdb-mcp-server bin entry.
• LICENSE-APACHE.txt — dual-licensed under Apache License 2.0 or MIT.
• InfluxDB 3 MCP docs — official InfluxData documentation for the server.

Repository facts confirmed at review time: official influxdata organization, default branch main, not archived, actively maintained (pushed within days of this submission), and released as @influxdata/influxdb3-mcp-server v1.3.0 on npm.

Features

• SQL querying — execute_query runs SQL against a database and supports multiple output formats.
• Data ingestion — write_line_protocol writes points using InfluxDB line protocol.
• Schema & discovery — list_databases, get_measurements, and get_measurement_schema inspect databases, measurements (tables), and column types.
• Database management — create_database, update_database (retention and, on Cloud Dedicated/Clustered, table/column limits), and delete_database (irreversible).
• Token management (Core/Enterprise) — create_admin_token, list_admin_tokens, create_resource_token, list_resource_tokens, delete_token, and regenerate_operator_token.
• Token management (Cloud Dedicated/Clustered) — cloud_list_database_tokens, cloud_get_database_token, cloud_create_database_token, cloud_update_database_token, and cloud_delete_database_token.
• Health & help — health_check reports connection/health status, and get_help returns troubleshooting guidance; load_database_context loads optional custom database context.
• MCP resources & prompts — read-only resources such as influx-config, influx-status, and influx-databases, plus a check-health prompt.
• Broad InfluxDB 3 coverage — one server works across Core, Enterprise, Cloud Serverless, Cloud Dedicated, and Clustered via INFLUX_DB_PRODUCT_TYPE.
• Multiple runtimes — run via npx (no install), a from-source Node build, or a Docker image you build from the repo (tagged mcp/influxdb).

Installation

Run the published package with npx (no clone required):

npx -y @influxdata/influxdb3-mcp-server

Add it to an MCP client (e.g. Claude Desktop's claude_desktop_config.json) for a Cloud Serverless instance (self-hosted Core/Enterprise instances use their own TLS instance URL and INFLUX_DB_PRODUCT_TYPE=core/enterprise):

{
  "mcpServers": {
    "influxdb": {
      "command": "npx",
      "args": ["-y", "@influxdata/influxdb3-mcp-server"],
      "env": {
        "INFLUX_DB_INSTANCE_URL": "https://us-east-1-1.aws.cloud2.influxdata.com",
        "INFLUX_DB_TOKEN": "<YOUR_INFLUXDB_TOKEN>",
        "INFLUX_DB_PRODUCT_TYPE": "cloud-serverless"
      }
    }
  }
}

Or run via Docker. The image is built from the repo first (docker compose build or npm run docker:build, which tags it mcp/influxdb), then invoked by the MCP client:

{
  "mcpServers": {
    "influxdb": {
      "command": "docker",
      "args": ["run", "--rm", "-i",
        "-e", "INFLUX_DB_INSTANCE_URL",
        "-e", "INFLUX_DB_TOKEN",
        "-e", "INFLUX_DB_PRODUCT_TYPE",
        "mcp/influxdb"],
      "env": {
        "INFLUX_DB_INSTANCE_URL": "https://us-east-1-1.aws.cloud2.influxdata.com",
        "INFLUX_DB_TOKEN": "<YOUR_INFLUXDB_TOKEN>",
        "INFLUX_DB_PRODUCT_TYPE": "cloud-serverless"
      }
    }
  }
}

For a Cloud Dedicated cluster, set INFLUX_DB_PRODUCT_TYPE=cloud-dedicated and INFLUX_DB_CLUSTER_ID, plus a token combination (database token for query/write, management token for database administration, or both for full functionality):

{
  "mcpServers": {
    "influxdb": {
      "command": "npx",
      "args": ["-y", "@influxdata/influxdb3-mcp-server"],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "cloud-dedicated",
        "INFLUX_DB_CLUSTER_ID": "<YOUR_CLUSTER_ID>",
        "INFLUX_DB_ACCOUNT_ID": "<YOUR_ACCOUNT_ID>",
        "INFLUX_DB_TOKEN": "<YOUR_DATABASE_TOKEN>",
        "INFLUX_DB_MANAGEMENT_TOKEN": "<YOUR_MANAGEMENT_TOKEN>"
      }
    }
  }
}

Key environment variables

* On Cloud Dedicated and Clustered, provide a database token, a management token (with INFLUX_DB_ACCOUNT_ID on Cloud Dedicated), or both, per the README's per-product options.

Use Cases

• Conversational analytics — ask natural-language questions and have the agent translate them into SQL against your InfluxDB 3 databases.
• Schema exploration — let an LLM enumerate databases, measurements, and column types before writing queries or application code.
• Operational data entry — write line-protocol points into a database from within an agentic workflow.
• Database administration — create, reconfigure (retention/limits), and remove databases through natural language.
• Token lifecycle management — mint, list, and revoke admin/resource/database tokens on Core, Enterprise, and Cloud clusters (with least-privilege scoping).

Safety and Privacy

• Write and delete capable. write_line_protocol, create_database/update_database, and the irreversible delete_database can change or destroy data; execute_query runs arbitrary SQL with the token's permissions.
• Credential-issuing tools. Token tools can create and revoke access — regenerate_operator_token is explicitly dangerous/irreversible. Treat created tokens as secrets and scope them tightly.
• Token is the boundary. Tool availability is not access control; use a least-privilege InfluxDB token (and a non-production instance where possible) when an agent acts autonomously.
• Data flows to the model. Query results, schemas, and database listings are returned to the LLM/MCP client and can include sensitive time-series tags or fields (device, user, or location identifiers).
• Credential hygiene. INFLUX_DB_TOKEN, instance URL, and cluster ID live in your client config's env block — keep it out of version control and restrict who can read it.

Duplicate Check

No existing entry in the directory matches this server. A search of all directory entries for "influxdb", "influx", and "influxdata" across slugs, titles, repository URLs, and install commands returned no dedicated entry, and there is no prior entry pointing at github.com/influxdata/influxdb3_mcp_server or the npm package @influxdata/influxdb3-mcp-server. This is therefore a net-new, non-duplicate entry.

• Source repositorygithub.com 2026-07-08T19:35:57+00:00
• Documentationdocs.influxdata.com
• Websiteinfluxdata.com
• Submitted by davion-knight2026-07-08