io.github.MukundaKatta/sqlfmt-mcp — MCP Server

This MCP server provides deterministic SQL formatting across 19 SQL dialects, including postgres, mysql, snowflake, and bigquery. It exposes tools to format SQL with consistent indentation and keyword casing so formatted output is repeatable for downstream diffing and code review.

🛠️ Key Features

Deterministic formatting across 19 dialects
Supports dialects such as postgres, mysql, snowflake, bigquery
Tool to format_sql with optional parameters like dialect, keyword_case, tab_width, and use_tabs

🚀 Use Cases

Canonicalizing SQL output for code-review agents and downstream tooling that diffs SQL
Reducing inconsistency in LLM-generated formatting across turns

⚡ Developer Benefits

list_dialects() helps confirm supported dialects before formatting
format_sql(...) returns formatted text + line count

⚠️ Limitations

The server is described as formatting SQL deterministically across 19 dialects (no additional capabilities are listed)

@mukundakatta/sqlfmt-mcp

MCP server for deterministic SQL formatting across 19 dialects.

Tools

format_sql(sql, dialect?, keyword_case?, tab_width?, use_tabs?) — format with the named dialect (default sql for ANSI). Returns formatted text + line count.
list_dialects() — return the list of supported dialects so an LLM can confirm before formatting.

Why this exists

LLMs format SQL inconsistently across turns: same query, different indentation, different keyword casing. Code-review agents and downstream tooling that diffs SQL want one canonical form. Wrapping the well-tested sql-formatter library here gives you that.

Supported dialects

sql, bigquery, db2, db2i, duckdb, hive, mariadb, mysql, n1ql, plsql, postgresql, redshift, singlestoredb, snowflake, spark, sqlite, tidb, transactsql, trino.

Install — Claude Desktop

json

{
  "mcpServers": {
    "sqlfmt": { "command": "npx", "args": ["-y", "@mukundakatta/sqlfmt-mcp"] }
  }
}

Same shape for Cursor / Cline / Windsurf / Zed.

License

MIT OR Apache-2.0.