Official YDB MCP server

YDB MCP

YDB's own official MCP server from the database's developers: parameterized SQL queries, execution plans, directory browsing, and a base for custom servers

MCP server OfficialEditors’ pickRussian stack

ydb-platform/ydb-mcpAuthor:YDB

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add mcp/ydb-mcp

Detects the agents on your machine, checks the risk and pins the version.

Other ways to install

This entry is high risk, so there is no one-click install. Review the code and add the config by hand.

Run in a terminal

claude mcp add --transport stdio ydb -- uvx ydb-mcp --ydb-endpoint grpc://localhost:2136 --ydb-database /local

Or add to the file .mcp.json, in the project

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcpServers key.

Add to the file ~/.cursor/mcp.json, for all projects

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcpServers key. For a single project, put the same block into .cursor/mcp.json.

Run in a terminal

code --add-mcp '{"name":"ydb","type":"stdio","command":"uvx","args":["ydb-mcp","--ydb-endpoint","grpc://localhost:2136","--ydb-database","/local"]}'

Or add to the file .vscode/mcp.json, in the project

{
  "servers": {
    "ydb": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the servers key.

Run in a terminal

codex mcp add ydb -- uvx ydb-mcp --ydb-endpoint grpc://localhost:2136 --ydb-database /local

Or add to the file ~/.codex/config.toml, for all projects

[mcp_servers.ydb]
command = "uvx"
args = ["ydb-mcp", "--ydb-endpoint", "grpc://localhost:2136", "--ydb-database", "/local"]

If the file already exists, append the block to the end.

Add to the file ~/.gemini/settings.json, for all projects

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcpServers key.

Add to the file ~/.config/devin/mcp_config.json, for all projects

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcpServers key. Legacy Cascade keeps the MCP config in ~/.codeium/windsurf/mcp_config.json.

Formerly Windsurf.

Add to the file cline_mcp_settings.json, for all projects

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcpServers key. Open the settings file in Cline: MCP Servers tab, Configure MCP Servers.

Add to the file .roo/mcp.json, in the project

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcpServers key.

A fork of Roo Code, same .roo folders.

Add to the file opencode.json, in the project

{
  "mcp": {
    "ydb": {
      "type": "local",
      "command": [
        "uvx",
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcp key.

Add to the file ~/.config/zed/settings.json, for all projects

{
  "context_servers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the context_servers key.

Add to the file .codeassistant/mcp.json, in the project

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint",
        "grpc://localhost:2136",
        "--ydb-database",
        "/local"
      ]
    }
  }
}

If the file already exists, add the server inside the mcpServers key.

You will need: uv

Checked against the repository on Sep 25, 2026, commit 1ee9df3.

Text for your agent

Add the server to your MCP client config with command uvx and args ydb-mcp --ydb-endpoint grpc://localhost:2136 --ydb-database /local; for a production database add --ydb-auth-mode with login/password, an access token, or a Yandex Cloud service account.

Other ways from the author

{
  "mcpServers": {
    "ydb": {
      "command": "uvx",
      "args": [
        "ydb-mcp",
        "--ydb-endpoint", "grpc://localhost:2136",
        "--ydb-database", "/local"
      ]
    }
  }
}

Runs via uvx with no system install, shown here with anonymous authentication.

This is third-party code. Review the repository files before installing.

What it does

The server gives an agent seven tools for the YDB distributed database: running a SQL query, running a parameterized query with JSON parameters, getting the execution plan for a plain or parameterized query, listing directory contents, describing a path in detail (tables, directories) and checking the connection status. It supports anonymous authentication, login/password, an access token, and Yandex Cloud service account authentication via the yandexcloud package. The YDBMCPServer class is designed to be subclassed: you can add your own tools on top of an established connection and disable the generic tools so the model only sees a fixed set of queries matching your business logic instead of arbitrary database access.

Who it is for. For developers and data engineers working with YDB who want to run queries and inspect the schema through an agent.

Good fit when

You want an official YDB MCP server maintained by its developers, not a third-party wrapper
You want a query's execution plan before running it against a production database
You want to restrict the agent to a fixed set of safe queries via your own server subclass

Not a fit when

You need protection from arbitrary SQL out of the box: the base server executes any query it is sent without restriction
You are working with a different database, not YDB

Example request

Show the execution plan for this YDB query and list the tables in the /local/orders directory

Limitations

The base server executes arbitrary SQL against the database, including data-changing statements, unless the toolset is explicitly narrowed via subclassing. Yandex Cloud service account authentication requires the extra yandexcloud package. It needs a reachable YDB grpc endpoint, local or cloud.

How to disable. Remove the ydb block from your MCP client config.

MCP

Transport: stdio
Authentication: API key

Security check

The base server executes arbitrary SQL, including data changes and deletions
Token or service account authentication grants broad database access unless the key's rights are restricted

README in short

The README describes YDB MCP as the official server for working with YDB databases from any MCP-capable LLM. It shows three install and run methods (uvx, pipx, pip), configuration examples for anonymous access, login/password, an access token and a Yandex Cloud service account, a list of seven tools, and a separate section on building custom MCP servers on top of the YDBMCPServer class, disabling generic tools for security and domain specificity.

FAQ

Can the agent be restricted to read-only access?

Yes, by subclassing YDBMCPServer, disabling the generic tools and adding your own with a limited set of queries.

Which authentication methods are supported?

Anonymous access, login/password, an access token and a Yandex Cloud service account.

Scientific Agent Skills

Editors’ pick

166 skills for scientific work: bioinformatics, cheminformatics, clinical data, geospatial analysis and 100+ databases

SkillMedium risk47K

MCP Toolbox for Databases

Official

Google's open-source MCP server for databases: ready tools for Postgres, MySQL, BigQuery, Spanner and more, plus custom tools in tools.yaml

MCP serverHigh risk16.5K

Hugging Face Skills

Editors’ pick

Official Hugging Face skills: Hub operations via the hf CLI, datasets, model training, Spaces, evals and deployment

SkillHigh risk11.1K

Flint

Official

A visualization language for agents and an MCP server: neat charts from a simple semantic spec

MCP serverMedium riskNo VPN needed4.3K

Install

What it does

Good fit when

Not a fit when

Example request

Limitations

MCP

Security check

README in short

FAQ

Related

Scientific Agent Skills

MCP Toolbox for Databases

Hugging Face Skills

Flint