create-bucket
Provision a new bucket under an organization so that subsequent write-data calls have a destination.
LABSCO SUMMARY
Writing telemetry and confirming it landed are both here, which is what makes it usable unattended.
query-data is described as the way to validate recently written data, so write, read back, correct is one loop in one conversation rather than a trip to another tool. The surface stops there: four tools, no deletes, no task or dashboard management, and no way to list what buckets already exist. This is a data-in and data-out server, and the provisioning tools are there so the first write has somewhere to go.
What it is
A four-tool InfluxDB client covering the write path, Flux queries, and the organization and bucket provisioning those depend on.
What you get
- Newline-delimited line protocol streamed into a named bucket, with timestamp precision set per call
- Flux queries run inside an organization, for inspecting measurement schemas, running aggregations, or checking what a write actually stored
- Bucket creation with a retentionPeriodSeconds, so a destination exists before the first write
- Organization creation, to keep separate projects or users on their own buckets and tokens
Requirements
INFLUXDB_URL for the instance, INFLUXDB_TOKEN with rights to write and to create buckets and organizations, and INFLUXDB_ORG as the organization queries run inside. MIT licensed.
Setup effort
One command plus a key — INFLUXDB_TOKEN=<YOUR_INFLUXDB_TOKEN> npx influxdb-mcp-server
USE CASES
01
Standing up storage for a new data source
Create the organization if this data should be isolated from what already exists, create the bucket with the retention you want, then start streaming line protocol into it. The bucket has to exist before the first write has anywhere to land.
02
Confirming a write landed the way you meant
Write the batch, then query it straight back with Flux to check the measurement, the tags and the timestamps came out as intended. Getting precision wrong is invisible until you read the points back.
03
Answering a trend question without leaving the conversation
Run the aggregation as a Flux query against the bucket and read the result inline, instead of exporting rows and working them somewhere else.
4 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
create-org
Create a brand-new organization to isolate users or projects before generating buckets and tokens.
query-data
Execute a Flux query inside an organization to inspect measurement schemas, run aggregations, or validate recently written data.
write-data
Stream newline-delimited line protocol records into a bucket. Use this after composing measurements so the LLM can insert real telemetry, optionally controlling timestamp precision.
BEFORE YOU INSTALL
What to know first
Tools you get
4 tools when we connected. Three of the four write: write-data inserts points, and create-bucket and create-org create objects. Names use hyphens, so it is write-data, not write_data.
Requires
- npx on PATH
- INFLUXDB_TOKEN in your environment
CONTEXT COST
Light Standing cost
Rated by the size of the 4 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
INFLUXDB_TOKEN=<YOUR_INFLUXDB_TOKEN> npx influxdb-mcp-serverMCP Clients
Claude Desktop — npx
CLAUDE DESKTOP — NPX · JSON
{
"mcpServers": {
"influxdb": {
"command": "npx",
"args": [
"influxdb-mcp-server"
],
"env": {
"INFLUXDB_TOKEN": "<YOUR_INFLUXDB_TOKEN>",
"INFLUXDB_URL": "http://localhost:8086",
"INFLUXDB_ORG": "<YOUR_INFLUXDB_ORG>"
}
}
}
}Claude Desktop — local build
CLAUDE DESKTOP — LOCAL BUILD · TERMINAL
git clone https://github.com/idoru/influxdb-mcp-server.git
cd influxdb-mcp-server
npm installPoint claude_desktop_config.json at src/index.js inside that folder, using its full path:
CLAUDE DESKTOP — LOCAL BUILD · JSON
{
"mcpServers": {
"influxdb": {
"command": "node",
"args": [
"<YOUR_PATH_TO_INFLUXDB_MCP_SERVER>/src/index.js"
],
"env": {
"INFLUXDB_TOKEN": "<YOUR_INFLUXDB_TOKEN>",
"INFLUXDB_URL": "http://localhost:8086",
"INFLUXDB_ORG": "<YOUR_INFLUXDB_ORG>"
}
}
}
}
