# InfluxData/influxdb3_mcp_server [Verified] [Health: Active]

**Category:** 🗄️ Databases  
**Repository:** https://github.com/influxdata/influxdb3_mcp_server  
**GitHub Stars:** 38  
**Views:** 58  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/influxdata-influxdb3-mcp-server

## Description
Official MCP server for InfluxDB 3 Core/Enterprise/Cloud Dedicated

## Tools
Capabilities this server exposes over MCP:

- **load_database_context** — Load optional custom database context and documentation
- **get_help** — Get help and troubleshooting guidance for InfluxDB operations
- **write_line_protocol** — Write data using InfluxDB line protocol
- **create_database** — Create a new database (with cloud-specific config options)
- **update_database** — Update database configuration (retention for all; maxTables/maxColumns for Cloud Dedicated/Clustered)
- **delete_database** — Delete a database by name (irreversible)
- **execute_query** — Deprecated since 2.0.0. Use `query_sql`. Retained for compatibility
- **query_sql** — Run bounded read-only SQL with structured response metadata
- **query_influxql** — Run bounded read-only InfluxQL with structured response metadata
- **get_measurements** — List all measurements (tables) in a database
- **get_measurement_schema** — Get schema (columns/types) for a measurement/table
- **list_tables** — List tables, also called measurements, in a database
- **describe_table** — Describe table schema with conservative column categories
- **investigate_database** — Run high-level read-only database discovery and sampling
- **create_admin_token** — Create a named admin token with an optional expiration
- **list_admin_tokens** — List all admin tokens (with optional filtering)
- **create_resource_token** — Create a resource token for specific DBs and permissions
- **list_resource_tokens** — List all resource tokens (with filtering and ordering)
- **delete_token** — Delete a token by name
- **regenerate_operator_token** — Regenerate the operator token (dangerous/irreversible)
- **cloud_list_database_tokens** — List all database tokens for Cloud-Dedicated/Clustered cluster
- **cloud_get_database_token** — Get details of a specific database token by ID
- **cloud_create_database_token** — Create a new database token for Cloud-Dedicated/Clustered cluster
- **cloud_update_database_token** — Update an existing database token
- **cloud_delete_database_token** — Delete a database token from Cloud-Dedicated/Clustered cluster
- **list_databases** — List all available databases in the instance
- **health_check** — Check InfluxDB connection and health status

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What InfluxData/influxdb3_mcp_server does

InfluxData/influxdb3_mcp_server exposes InfluxDB 3 operations through the Model Context Protocol. It supports InfluxDB 3 Core and Enterprise, along with Cloud Dedicated, Cloud Clustered, and Cloud Serverless deployments. The tool set covers data access, schema inspection, database lifecycle operations, and token administration.

Agents can list databases and tables, inspect measurement schemas, sample an unfamiliar database, and check connection health. They can query with SQL or InfluxQL, write line protocol, and load custom database context or operational help. Administrative tools create, update, and delete databases, while token tools create, list, delete, or regenerate credentials depending on the deployment.

## How it works

The server receives MCP tool calls and sends requests to the configured InfluxDB instance or cloud service. Core, Enterprise, and Cloud Serverless connections use an instance URL and token; Cloud Dedicated and Cloud Clustered connections use a cluster ID and tokens. The supplied credential is sent as a Bearer token, so database visibility and permitted actions are controlled by that token.

`query_sql` and `query_influxql` perform bounded, read-only queries and return structured response metadata. Results can include row counts, truncation status, warnings, and correlation identifiers. Query arguments use `database` and `query`; calls that provide both aliases with conflicting values fail with a structured argument error.

Discovery tools distinguish known values from unavailable ones. `investigate_database` preserves results for tables that succeed even if another table fails, and samples a 24-hour window by default. On Core and Enterprise, schema investigation can use authoritative server schema and internal catalog data when the credential permits access.

Set `INFLUX_MCP_TOOL_PROFILE=readonly` to expose only exploration and read-query tools. This profile excludes data writes, database administration, token management, and host operations.

## Setup and configuration

Node.js 20.11 or newer is required for npm or npx usage. Docker can be used for a container-based setup. Local development uses pnpm, with the version pinned in `package.json`; Corepack can activate pnpm with Node.js 20 or 22.

Provide an InfluxDB URL and token for Core, Enterprise, or Cloud Serverless. Cloud Dedicated and Cloud Clustered deployments use a cluster ID and the relevant tokens. Set `INFLUX_DB_TOKEN` to an administrator token, scoped token, or Enterprise user-authentication JWT. The server passes this value through without inspecting its permissions.

## Tools and capabilities

- Run bounded read-only SQL with `query_sql` or InfluxQL with `query_influxql`; `execute_query` remains only for compatibility and is deprecated.
- Explore databases and schemas with `list_databases`, `list_tables`, `get_measurements`, `describe_table`, `get_measurement_schema`, and `investigate_database`.
- Write line protocol and manage database configuration, including retention settings and applicable cloud limits for tables and columns.
- Create, list, delete, and regenerate tokens, with separate database-token operations for Cloud Dedicated and Cloud Clustered deployments.
- Load database-specific context, retrieve troubleshooting guidance, and check health with `health_check`.

## Limitations and notes

Read results are bounded, and a complete answer should account for metadata such as completeness, caps, and warnings. Queries that exceed server memory require narrower requests. An Enterprise ingest-only endpoint cannot serve query operations. A 403 response is reported as `not_found_or_denied`, because a credential without access cannot distinguish an absent database from a forbidden one.

Database deletion and operator-token regeneration are irreversible operations. `EXPLAIN ANALYZE` runs within the same query bounds and adds a diagnostic warning. Cardinality estimates are not supported. When authoritative deletion metadata is unavailable, a timestamp-shaped table suffix may produce a `possibly_deleted` advisory while the table remains listed.

_Full upstream README: https://allmcps.com/mcp/influxdata-influxdb3-mcp-server/readme_