Dremio MCP Server

Name: Dremio MCP Server
Author: Dremio

Dremio's own MCP server for browsing lakehouse tables, running SQL and analysing cluster usage

Official · Dremio Python Local · stdio Databases Data and analytics

GitHub Website

29 popularity score ★ 53

Connect your client

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 DREMIOAI_DREMIO__URI, DREMIOAI_DREMIO__PAT. Replace each <NAME> with your value.

FirstTerminal

git clone https://github.com/dremio/dremio-mcp.git <REPO_DIR>
cd <REPO_DIR>
uv run dremio-mcp-server config create dremioai --uri <DREMIO_URI> --pat <DREMIO_PAT>

Claude CodeTerminal

claude mcp add dremio -e DREMIOAI_DREMIO__URI=<DREMIOAI_DREMIO__URI> -e DREMIOAI_DREMIO__PAT=<DREMIOAI_DREMIO__PAT> -- uv run --directory '<REPO_DIR>' dremio-mcp-server run

Claude Desktopclaude_desktop_config.json

{
  "mcpServers": {
    "dremio": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "<REPO_DIR>",
        "dremio-mcp-server",
        "run"
      ],
      "env": {
        "DREMIOAI_DREMIO__URI": "<DREMIOAI_DREMIO__URI>",
        "DREMIOAI_DREMIO__PAT": "<DREMIOAI_DREMIO__PAT>"
      }
    }
  }
}

Cursor~/.cursor/mcp.json

{
  "mcpServers": {
    "dremio": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "<REPO_DIR>",
        "dremio-mcp-server",
        "run"
      ],
      "env": {
        "DREMIOAI_DREMIO__URI": "<DREMIOAI_DREMIO__URI>",
        "DREMIOAI_DREMIO__PAT": "<DREMIOAI_DREMIO__PAT>"
      }
    }
  }
}

VS Code.vscode/mcp.json

{
  "servers": {
    "dremio": {
      "type": "stdio",
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "<REPO_DIR>",
        "dremio-mcp-server",
        "run"
      ],
      "env": {
        "DREMIOAI_DREMIO__URI": "<DREMIOAI_DREMIO__URI>",
        "DREMIOAI_DREMIO__PAT": "<DREMIOAI_DREMIO__PAT>"
      }
    }
  }
}

Codex~/.codex/config.toml

[mcp_servers.dremio]
command = "uv"
args = ["run", "--directory", "<REPO_DIR>", "dremio-mcp-server", "run"]

[mcp_servers.dremio.env]
DREMIOAI_DREMIO__URI = "<DREMIOAI_DREMIO__URI>"
DREMIOAI_DREMIO__PAT = "<DREMIOAI_DREMIO__PAT>"

Gemini CLITerminal

gemini mcp add --scope user -e DREMIOAI_DREMIO__URI=<DREMIOAI_DREMIO__URI> -e DREMIOAI_DREMIO__PAT=<DREMIOAI_DREMIO__PAT> dremio uv -- run --directory '<REPO_DIR>' dremio-mcp-server run

Settings 4

`DREMIOAI_DREMIO__URI`	Required: Dremio API endpoint: https://api.dremio.cloud (or prod), https://api.eu.dremio.cloud (prodemea), or your coordinator URL; can be set in config.yaml instead.
`DREMIOAI_DREMIO__PAT`	Required: Dremio personal access token; can also be set in config.yaml, or from a file with an @ prefix.
`DREMIOAI_TOOLS__SERVER_MODE`	Optional: Tool sets to enable: FOR_DATA_PATTERNS, FOR_SELF, FOR_PROMETHEUS, EXPERIMENTAL, comma-separated.
`JSON_LOGGING`	Optional: Set to 1 for structured JSON log output.

What it does

Dremio MCP Server is a Python program from Dremio that connects an MCP client to a Dremio lakehouse, either Dremio Cloud (US or EMEA) or a self-managed Dremio cluster. Through it an assistant can look up table and view schemas, read table descriptions and lineage, search the catalog, and run SQL against the data.

What it offers depends on the mode set in its config file. FOR_DATA_PATTERNS is the everyday mode for exploring tables and finding patterns in data. FOR_SELF turns it on Dremio itself, with access to system tables and usage reports by engine or project. FOR_PROMETHEUS adds tools that query a Prometheus server watching the cluster. Modes can be combined, and a further set of tools passes calls through to extra tools that a Dremio server exposes, when that is enabled.

It runs in two ways. Locally, it is cloned from GitHub and started with uv as a stdio server beside a desktop client such as Claude Desktop, and a built-in config command writes both its own settings file and the Claude config. For teams, the repository ships a Dockerfile and a Helm chart that run it on Kubernetes as a streaming-HTTP service with OAuth sign-in. The repository says it is open source and not part of Dremio product support.

What it needs

Runtime	Python 3.11+

A Dremio Cloud project or a self-managed Dremio cluster, with a personal access token (plus the project ID for Dremio Cloud). Python 3.11 or later and the uv package manager, with a local clone of the repository. The remote option needs Docker, Kubernetes and Helm; Prometheus tools need a Prometheus endpoint.

Limits

SQL is limited to SELECT-style queries unless allow_dml is switched on; results are capped at a size the server sets, and an oversized result returns an error asking for a narrower query. Not published as a package: it runs from a cloned repository, and the Docker image is built yourself. The repository states it is not covered by Dremio product support. Some tools sit behind an experimental flag, and semantic-layer tools return nothing unless that feature is configured.

Who it suits

Suits Dremio users who want an assistant to explore their lakehouse catalog and query data with SQL, and Dremio administrators who want help reading system tables, job usage and Prometheus metrics. Read-only by default, so it fits analysis work; turn on allow_dml only when views or other changes are wanted. Teams can host it centrally on Kubernetes with OAuth instead of handing out personal tokens. Not useful without a Dremio deployment; for other warehouses pick a server made for that database.

Tools 14

RunSqlQuery: Run a SQL query on Dremio; non-SELECT statements only when allow_dml is on
GetSchemaOfTable: Return the column schema of one or more tables
GetDescriptionOfTableOrSchema: Return the description recorded for a table or schema
GetTableOrViewLineage: Show which tables and views a table or view comes from
GetUsefulSystemTableNames: List the Dremio system tables most worth querying
SearchTableAndViews: Search the catalog for tables and views matching a query
BuildUsageReport: Summarise cluster usage grouped by engine or project
GetRelevantMetrics: List Prometheus metrics that relate to the Dremio cluster
GetMetricSchema: Show the labels and sample values of a Prometheus metric
RunPromQL: Run a PromQL query against the configured Prometheus server
SearchMetrics: Find semantic-layer metrics from a plain-language description
GetTableRelationships: Fetch semantic-layer relationship data for a table or view
DiscoverDynamicTools: List the extra tools the Dremio server makes available
CallDynamicTool: Call one of those server-provided tools by name with arguments

Popularity

29 popularity score

★ 53

GitHub stars	53	20.8 points
Official server	Yes	8 points

Measured 2026-11-07. How the score works.

Where it is published

Code	dremio/dremio-mcp ★ 53

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 (GitHub)
Last code change	2026-10-30

Connects over the network to the Dremio API endpoint you set (Dremio Cloud or your own coordinator) and, if configured, to a Prometheus server. Signs in with a Dremio personal access token kept in a local YAML config file or environment variable, or with OAuth in the Helm deployment; the docs mark the PAT setup as not for production in that deployment. RunSqlQuery blocks DML by default, but with allow_dml enabled it can change data and create views, and CallDynamicTool runs whatever extra tools the Dremio server offers. Runs on your machine, or on your own Kubernetes cluster. Writes log files to a per-user log directory by default.

Over time

GitHub stars 53 +0

5353

14 Oct 20267 Nov 2026

Popularity score 29 +0

290

14 Oct 20267 Nov 2026

What is rising across the directory

Sources

Page checked 2026-10-14.

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

[![Dremio MCP Server on EveryMCPServer](https://everymcpserver.com/badge/dremio-mcp-server.svg)](https://everymcpserver.com/servers/dremio-mcp-server)

Connect your client

Add in one click

Or add it by command or config

Settings 4

What it does

What it needs

Limits

Who it suits

Tools 14

Similar servers

dbt MCP Server

ClickHouse MCP Server

MySQL MCP Server

Redash MCP Server

Power BI Authoring MCP Server

MCPQL