Format SQL across 19 dialects (PostgreSQL, MySQL, BigQuery, Snowflake, etc.) with keyword casing and indentation control via @mukundakatta/sqlfmt-mcp

@mukundakatta/sqlfmt-mcp v latest — SQL formatter with 19 dialect support

Install: npm install @mukundakatta/sqlfmt-mcp Run: node node_modules/@mukundakatta/sqlfmt-mcp/src/index.js (stdio transport)

Tools (2)

1. `format_sql` ({sql, dialect?, tab_width?, use_tabs?, keyword_case?}) — Format a SQL string
2. dialect: one of 19 values (default "sql"): sql, bigquery, db2, db2i, duckdb, hive, mariadb, mysql, n1ql, plsql, postgresql, redshift, singlestoredb, snowflake, spark, sqlite, tidb, transactsql, trino
3. tab_width: 1-8 (default 2)
4. use_tabs: boolean (default false)
5. keyword_case: "preserve" | "upper" | "lower" (default "upper")
6. Returns: {formatted, dialect, line_count}

1. `list_dialects` ({}) — Returns {dialects: [...]} array of 19 supported dialects

Verified execution trace (10 calls, 10 success, 0 failures)

Simple SELECT (default dialect, keyword_case=upper):

→ format_sql({sql: "select id, name, email from users where status='active' and created_at > '2024-01-01' order by name asc limit 10"})
← {formatted: "SELECT\n  id,\n  name,\n  email\nFROM\n  users\nWHERE\n  status = 'active'\n  AND created_at > '2024-01-01'\nORDER BY\n  name ASC\nLIMIT\n  10", dialect: "sql", line_count: 13}

PostgreSQL with array_agg:

→ format_sql({sql: "SELECT u.id, u.name, array_agg(o.id) as order_ids FROM users u LEFT JOIN orders o ON u.id = o.user_id WHERE u.status = 'active' GROUP BY u.id, u.name HAVING count(o.id) > 5", dialect: "postgresql"})
← {formatted: "SELECT\n  u.id,\n  u.name,\n  array_agg(o.id) AS order_ids\nFROM\n  users u\n  LEFT JOIN orders o ON u.id = o.user_id\nWHERE\n  u.status = 'active'\nGROUP BY\n  u.id,\n  u.name\nHAVING\n  count(o.id) > 5", dialect: "postgresql", line_count: 14}

INSERT with tab_width=4:

→ format_sql({sql: "insert into users (name, email, role) values ('Alice', '[email protected]', 'admin'), ('Bob', '[email protected]', 'user')", tab_width: 4})
← {formatted: "INSERT INTO\n    users (name, email, role)\nVALUES\n    ('Alice', '[email protected]', 'admin'),\n    ('Bob', '[email protected]', 'user')", dialect: "sql", line_count: 5}

MySQL with backtick identifiers:

→ format_sql({sql: "SELECT `user`.`id`, `user`.`name` FROM `user` WHERE `user`.`deleted_at` IS NULL AND `user`.`role` IN ('admin', 'editor') ORDER BY `user`.`created_at` DESC", dialect: "mysql"})
← dialect: "mysql", line_count: 10 — backtick identifiers preserved correctly

Complex CTE:

→ format_sql({sql: "WITH active_users AS (SELECT id, name FROM users WHERE status = 'active'), user_orders AS (...) SELECT au.name, uo.order_count, uo.total_spent FROM active_users au JOIN user_orders uo ON au.id = uo.user_id WHERE uo.total_spent > 1000 ORDER BY uo.total_spent DESC"})
← Properly indented WITH/AS blocks, nested SELECT inside CTEs indented 4 spaces, main query at root level

keyword_case=lower:

→ format_sql({sql: "SELECT * FROM products WHERE price BETWEEN 10 AND 100", keyword_case: "lower"})
← "select\n  *\nfrom\n  products\nwhere\n  price between 10 and 100\n  and category in ('electronics', 'books')\n  and name like '%pro%'"

BigQuery COUNTIF:

→ format_sql({sql: "SELECT DATE_TRUNC(created_at, MONTH) as month, COUNTIF(status = 'completed') as completed FROM `project.dataset.orders` GROUP BY 1 ORDER BY 1", dialect: "bigquery"})
← Preserves backtick-quoted project.dataset.table, COUNTIF recognized, GROUP BY 1 preserved

use_tabs=true:

→ format_sql({sql: "select a, b, c from t where a > 1", use_tabs: true})
← "SELECT\n\ta,\n\tb,\n\tc\nFROM\n\tt\nWHERE\n\ta > 1" — tab characters used instead of spaces

Performance

• p50: 5ms, range: 0-26ms
• First 2-3 calls ~15-26ms (JIT warmup), subsequent calls 1-5ms
• list_dialects returns instantly (0ms)

Key gotchas

• Default keyword_case is "upper" — keywords like SELECT, FROM, WHERE are uppercased by default; use keyword_case: "lower" or `"p