InfluxDB 3 MCP Server

Name: InfluxDB 3 MCP Server
Author: InfluxData

Query, write to and administer InfluxDB 3 time-series databases from an MCP client

Official · InfluxData TypeScript Local · stdio Databases DevOps and monitoring

GitHub npm Website

49 popularity score ★ 38 ↓ 593/wk

Connect your client

Local (npm)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio Add to Goose

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs INFLUX_DB_PRODUCT_TYPE, INFLUX_DB_INSTANCE_URL, INFLUX_DB_TOKEN. Replace each <NAME> with your value.

Claude CodeTerminal

claude mcp add influxdb-3 -e INFLUX_DB_PRODUCT_TYPE=<INFLUX_DB_PRODUCT_TYPE> -e INFLUX_DB_INSTANCE_URL=<INFLUX_DB_INSTANCE_URL> -e INFLUX_DB_TOKEN=<INFLUX_DB_TOKEN> -- npx -y @influxdata/influxdb3-mcp-server

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "influxdb-3": {
      "command": "npx",
      "args": [
        "-y",
        "@influxdata/influxdb3-mcp-server"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "influxdb-3": {
      "command": "npx",
      "args": [
        "-y",
        "@influxdata/influxdb3-mcp-server"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "influxdb-3": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@influxdata/influxdb3-mcp-server"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.influxdb-3]
command = "npx"
args = ["-y", "@influxdata/influxdb3-mcp-server"]

[mcp_servers.influxdb-3.env]
INFLUX_DB_PRODUCT_TYPE = "<INFLUX_DB_PRODUCT_TYPE>"
INFLUX_DB_INSTANCE_URL = "<INFLUX_DB_INSTANCE_URL>"
INFLUX_DB_TOKEN = "<INFLUX_DB_TOKEN>"

Gemini CLITerminal

gemini mcp add --scope user -e INFLUX_DB_PRODUCT_TYPE=<INFLUX_DB_PRODUCT_TYPE> -e INFLUX_DB_INSTANCE_URL=<INFLUX_DB_INSTANCE_URL> -e INFLUX_DB_TOKEN=<INFLUX_DB_TOKEN> influxdb-3 npx -- -y @influxdata/influxdb3-mcp-server

Local (container)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio Add to Goose

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs INFLUX_DB_PRODUCT_TYPE, INFLUX_DB_INSTANCE_URL, INFLUX_DB_TOKEN. Replace each <NAME> with your value.

Claude CodeTerminal

claude mcp add influxdb-3 -e INFLUX_DB_PRODUCT_TYPE=<INFLUX_DB_PRODUCT_TYPE> -e INFLUX_DB_INSTANCE_URL=<INFLUX_DB_INSTANCE_URL> -e INFLUX_DB_TOKEN=<INFLUX_DB_TOKEN> -- docker run --rm -i -e INFLUX_DB_INSTANCE_URL -e INFLUX_DB_TOKEN -e INFLUX_DB_PRODUCT_TYPE influxdata/influxdb3-mcp-server

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "influxdb-3": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "INFLUX_DB_INSTANCE_URL",
        "-e",
        "INFLUX_DB_TOKEN",
        "-e",
        "INFLUX_DB_PRODUCT_TYPE",
        "influxdata/influxdb3-mcp-server"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "influxdb-3": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "INFLUX_DB_INSTANCE_URL",
        "-e",
        "INFLUX_DB_TOKEN",
        "-e",
        "INFLUX_DB_PRODUCT_TYPE",
        "influxdata/influxdb3-mcp-server"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "influxdb-3": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "INFLUX_DB_INSTANCE_URL",
        "-e",
        "INFLUX_DB_TOKEN",
        "-e",
        "INFLUX_DB_PRODUCT_TYPE",
        "influxdata/influxdb3-mcp-server"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.influxdb-3]
command = "docker"
args = ["run", "--rm", "-i", "-e", "INFLUX_DB_INSTANCE_URL", "-e", "INFLUX_DB_TOKEN", "-e", "INFLUX_DB_PRODUCT_TYPE", "influxdata/influxdb3-mcp-server"]

[mcp_servers.influxdb-3.env]
INFLUX_DB_PRODUCT_TYPE = "<INFLUX_DB_PRODUCT_TYPE>"
INFLUX_DB_INSTANCE_URL = "<INFLUX_DB_INSTANCE_URL>"
INFLUX_DB_TOKEN = "<INFLUX_DB_TOKEN>"

Gemini CLITerminal

gemini mcp add --scope user -e INFLUX_DB_PRODUCT_TYPE=<INFLUX_DB_PRODUCT_TYPE> -e INFLUX_DB_INSTANCE_URL=<INFLUX_DB_INSTANCE_URL> -e INFLUX_DB_TOKEN=<INFLUX_DB_TOKEN> influxdb-3 docker -- run --rm -i -e INFLUX_DB_INSTANCE_URL -e INFLUX_DB_TOKEN -e INFLUX_DB_PRODUCT_TYPE influxdata/influxdb3-mcp-server

Local (from source)

Add in one click

Add to Cursor Add to VS Code Add to LM Studio

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs INFLUX_DB_PRODUCT_TYPE, INFLUX_DB_INSTANCE_URL, INFLUX_DB_TOKEN. Replace each <NAME> with your value.

FirstTerminal

git clone https://github.com/influxdata/influxdb3_mcp_server.git
cd influxdb3_mcp_server
corepack enable
pnpm install
pnpm run build

Claude CodeTerminal

claude mcp add influxdb-3 -e INFLUX_DB_PRODUCT_TYPE=<INFLUX_DB_PRODUCT_TYPE> -e INFLUX_DB_INSTANCE_URL=<INFLUX_DB_INSTANCE_URL> -e INFLUX_DB_TOKEN=<INFLUX_DB_TOKEN> -- node '<ABSOLUTE_PATH_TO_CLONE>/build/index.js'

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "influxdb-3": {
      "command": "node",
      "args": [
        "<ABSOLUTE_PATH_TO_CLONE>/build/index.js"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "influxdb-3": {
      "command": "node",
      "args": [
        "<ABSOLUTE_PATH_TO_CLONE>/build/index.js"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "influxdb-3": {
      "type": "stdio",
      "command": "node",
      "args": [
        "<ABSOLUTE_PATH_TO_CLONE>/build/index.js"
      ],
      "env": {
        "INFLUX_DB_PRODUCT_TYPE": "<INFLUX_DB_PRODUCT_TYPE>",
        "INFLUX_DB_INSTANCE_URL": "<INFLUX_DB_INSTANCE_URL>",
        "INFLUX_DB_TOKEN": "<INFLUX_DB_TOKEN>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.influxdb-3]
command = "node"
args = ["<ABSOLUTE_PATH_TO_CLONE>/build/index.js"]

[mcp_servers.influxdb-3.env]
INFLUX_DB_PRODUCT_TYPE = "<INFLUX_DB_PRODUCT_TYPE>"
INFLUX_DB_INSTANCE_URL = "<INFLUX_DB_INSTANCE_URL>"
INFLUX_DB_TOKEN = "<INFLUX_DB_TOKEN>"

Gemini CLITerminal

gemini mcp add --scope user -e INFLUX_DB_PRODUCT_TYPE=<INFLUX_DB_PRODUCT_TYPE> -e INFLUX_DB_INSTANCE_URL=<INFLUX_DB_INSTANCE_URL> -e INFLUX_DB_TOKEN=<INFLUX_DB_TOKEN> influxdb-3 node -- '<ABSOLUTE_PATH_TO_CLONE>/build/index.js'

Settings 10

Set together: credentials
`INFLUX_DB_PRODUCT_TYPE`	Required: core, enterprise, cloud-serverless, cloud-dedicated or clustered; the server refuses to start without a valid value.
`INFLUX_DB_INSTANCE_URL`	Required: Address of the instance, e.g. http://localhost:8181/; checked at start for Core, Enterprise and Cloud Serverless, and needed for Clustered.
`INFLUX_MCP_TOOL_PROFILE`	Optional: Set to readonly to expose only the nine read-only tools; any other value keeps all tools. Default `operator`.
`INFLUX_DB_TOKEN` secret	Required: Admin, scoped or database token, sent as a Bearer header; on Dedicated/Clustered INFLUX_DB_MANAGEMENT_TOKEN can stand in. INFLUX_DB_DATABASE_TOKEN is read as a fallback.
`INFLUX_DB_MANAGEMENT_TOKEN` secret	Optional: Management token for the cloud management API. For database and token management on Cloud Dedicated or Clustered.
Set together: cloud-dedicated
`INFLUX_DB_CLUSTER_ID`	Optional: Cluster ID; required for Cloud Dedicated.
`INFLUX_DB_ACCOUNT_ID`	Optional: Account ID; needed with the management token on Cloud Dedicated.
Set together: logging
`MCP_LOG_TOOL_CALLS`	Optional: Set to false to stop the tool-call log; on by default.
`MCP_LOG_BACKEND`	Optional: file writes the log to MCP_LOG_FILE (falls back to stderr if unset); stdout is also accepted but shares the MCP channel. Default `stderr`.
`MCP_LOG_FILE`	Optional: Path of the log file when MCP_LOG_BACKEND=file.

In a container, a setting goes in with -e NAME=value after docker run.

What it does

InfluxData's own MCP server for InfluxDB 3. You point it at one instance of Core, Enterprise, Cloud Serverless, Cloud Dedicated or Clustered with INFLUX_DB_PRODUCT_TYPE, and it runs on your machine over stdio, from npm or as a container.

An assistant can use it to list databases and tables, look at column schemas, run SQL or InfluxQL, and write points in line protocol. It can also manage databases and tokens. The token tools depend on the product: admin tokens on Core and Enterprise, resource tokens on Enterprise only, and the cloud_* token tools on Cloud Dedicated and Clustered. There are four resources, three prompts, and a load_database_context tool that hands back a context file you write yourself to describe your data.

Version 2.0 changed the shape of query results and health-check output, so a client that parses those needs updating. execute_query is deprecated since 2.0.0 and stays only for older clients. Use query_sql instead, which returns 1000 rows by default and caps maxRows at 5000 (src/tools/categories/query.tools.ts).

All tools are exposed unless INFLUX_MCP_TOOL_PROFILE=readonly is set. That profile keeps nine of them: the context loader, help, health, database and table listing, table description, the two bounded query tools and a discovery pass. The config at v2.0.0 also takes readwrite and operator as values, but only readonly filters anything (src/tools/index.ts). Each tool call is written as a JSON line to stderr unless MCP_LOG_TOOL_CALLS=false. Tokens and query text are left out of that log.

What it needs

Runtime	Node.js 22+Docker
Runs on	macOS, Windows, Linux
InfluxDB 3	Your own instanceCore, Enterprise or Clustered you run, or a Cloud Serverless or Cloud Dedicated account, plus a token for it

A reachable InfluxDB 3 instance and a token for it. Cloud Dedicated also needs its cluster ID, and an account ID plus management token for database management. The README asks for Node.js 22.23.4+, 24.22.1+ or 26.12.1+ for npx, because older releases trust root certificates tied to compromised eMudhra/emSign CAs. package.json still allows Node 20.11. Docker is the alternative, and its image ships a current Node 24.

Limits

Not covered Works with InfluxDB 3 products only (Core, Enterprise, Cloud Serverless, Cloud Dedicated, Clustered)
Not covered On Cloud Dedicated and Clustered a database token alone gives query and write but no management, a management token alone the reverse
Size cap query_sql and query_influxql return 1000 rows by default, 5000 at most with maxRows
Limit execute_query is deprecated since 2.0.0 and returns JSON only on Cloud Dedicated, Cloud Serverless and Clustered
Limit 2.0 changed query result envelopes, health-check output and argument errors, so clients parsing 1.x output must be updated
Not covered investigate_database samples the last 24 hours by default (sampleWindowHours, up to 8760). Cardinality estimates are not supported yet
Limit The README's npx section still says it will not work until published, and its Docker examples use local image names, not influxdata/influxdb3-mcp-server

Works with InfluxDB 3 products only. On Cloud Dedicated and Clustered, a database token alone gives query and write tools but no database management, and a management token alone gives management but no queries; both are needed for everything. query_sql and query_influxql return 1000 rows by default, 5000 at most; execute_query returns JSON only on Cloud Dedicated, Cloud Serverless and Clustered. The README's npx section still says it will not work until the package is published, which is out of date; its Docker examples use locally built image names (mcp/influxdb, influxdb-mcp-server), while released images are on Docker Hub as influxdata/influxdb3-mcp-server.

Who it suits

Suits teams running InfluxDB 3 who want an assistant to explore schemas, answer questions in SQL or InfluxQL, or work out why a dashboard query broke. Use the readonly profile when it should change nothing. The full profile can also delete databases and manage tokens, so give it a token scoped to what you are willing to let it do. It covers InfluxDB 3 products only, so for InfluxDB 1.x or 2.x look elsewhere.

Tools 27

15 only read, 8 change things, 4 can delete or overwrite. A ? marks an optional parameter.

load_database_context: Return the contents of your own context file describing the databases
get_help: Give guidance and troubleshooting tips for InfluxDB tasks
write_line_protocol changes: Write points to a database in line protocol
create_database changes: Create a database, with cloud-specific options where they apply
update_database changes: Change a database's retention, and table or column caps on Dedicated/Clustered
delete_database deletes: Delete a database by name; cannot be undone
execute_query changes: Deprecated since 2.0.0: run any SQL query, results in a chosen formatdatabase: stringquery: stringformat?: enum
query_sql: Run a read-only, row-capped SQL query with structured metadatadb?: stringq?: stringdatabase?: stringquery?: stringparams?: anyformat?: enummaxRows?: integertimeoutMs?: integer
query_influxql: Run a read-only, row-capped InfluxQL query with structured metadatadb?: stringq?: stringdatabase?: stringquery?: stringparams?: anyformat?: enummaxRows?: integertimeoutMs?: integer
get_measurements: List the measurements (tables) in a databasedatabase: string
get_measurement_schema: Show the columns and types of one measurementdatabase: stringmeasurement: string
list_tables: List the tables in a databasedb: string
describe_table: Describe a table's columns, marking unclear tag/field roles as unknowndb: stringtable: string
investigate_database: Survey a database's tables and schemas, optionally with recent sample rowsdb: stringincludeSamples?: booleanmaxTables?: integersampleWindowHours?: integersampleRowsPerTable?: integer
create_admin_token changes: Create a named admin token, optionally expiring (Core/Enterprise)
list_admin_tokens: List admin tokens with optional filters (Core/Enterprise)
create_resource_token changes: Create a token limited to chosen databases and permissions (Enterprise)
list_resource_tokens: List resource tokens with filtering and ordering (Enterprise)
delete_token deletes: Delete a token by name (Core/Enterprise)
regenerate_operator_token deletes: Replace the operator token; the old one stops working (Core/Enterprise)
cloud_list_database_tokens: List database tokens on a Dedicated or Clustered cluster
cloud_get_database_token: Show one database token by ID
cloud_create_database_token changes: Create a database token on a Dedicated or Clustered cluster
cloud_update_database_token changes: Change an existing database token
cloud_delete_database_token deletes: Delete a database token from the cluster
list_databases: List the databases on the instance
health_check: Check the connection to InfluxDB and report health and ping results

Resources 4

influx://config: The connection settings in use, read-only
influx://status: Current connection and health status
influx://databases: The databases on the instance
influx://context: Your own context file describing the data

Prompts 3

list-databases: Ask for a list of the databases
check-health: Ask for a health check of the instance
load-context: Load your context file into the conversation

Popularity

49 popularity score, on our list (the bar is 36)

★ 38 ↓ 593/wk

GitHub stars	38	19.1 points
npm downloads, last week	593	22.2 points
Official server	Yes	8 points

Measured 2026-11-08. How the score works.

Where it is published

Code	influxdata/influxdb3_mcp_server ★ 38
npm	@influxdata/influxdb3-mcp-server 2.0.1, 2026-10-30, the one the command runs
Docker Hub	influxdata/influxdb3-mcp-server

What it can reach

Runs	On your machine
Reaches	A database, The internet
Changes things	Yes: it can write, send, create or delete
Sign-in	An API key or token
Licence	(Apache-2.0 OR MIT) (npm)
Usage data	Sends none Tool-call records (tool name, IDs, duration, database, row count, outcome) go only to stderr or a local file
Last code change	2026-10-29

It runs on your machine. Requests go over HTTP(S) to the InfluxDB instance you configure, with your token in a Bearer header. With the default profile the assistant can write data and drop databases. It can also mint or delete tokens, including regenerating the Core/Enterprise operator token, after which the old operator token no longer works. Setting INFLUX_MCP_TOOL_PROFILE=readonly cuts this to nine tools that list, describe and run bounded queries; the deprecated execute_query, which accepts any SQL, is left out. Whatever the profile, the token's permissions are the real limit. A local log of tool calls goes to stderr by default and omits tokens and query text.

Over time

GitHub stars 38 +0

3838

14 Oct 20268 Nov 2026

npm downloads a week 593 +549 (+1248 %)

1.3k24

25 May 20268 Nov 2026

Before the site measured it, read from npm's daily download counts.

Popularity score 49 +2 (+4 %)

5242

14 Oct 20268 Nov 2026

What is rising across the directory

Sources

Page checked 2026-10-30.

Badge for your README

Shows this server's rank in its category, and stars for a server that runs on your machine. It updates every hour.

MarkdownREADME.md

[![InfluxDB 3 MCP Server on EveryMCPServer](https://everymcpserver.com/badge/influxdb-3.svg)](https://everymcpserver.com/servers/influxdb-3)

Connect your client

Add in one click

Or add it by command or config

Add in one click

Or add it by command or config

Add in one click

Or add it by command or config

Settings 10

What it does

What it needs

Limits

Who it suits

Tools 27

Resources 4

Prompts 3

Similar servers

BetterDB MCP server

Confluent MCP Server

Elasticsearch MCP Server

Confluent Cloud managed MCP server

MySQL MCP Server

dbt MCP Server