Format and pretty-print SQL across 19 dialects (PostgreSQL, MySQL, BigQuery, Snowflake, etc.) via @mukundakatta/sqlfmt-mcp

@mukundakatta/sqlfmt-mcp v0.1.0 — SQL formatting via MCP

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

Tools (2)

19 supported dialects

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

Key gotchas

• Tool name is `format_sql` NOT `format` — format returns unknown tool error silently
• Default `keyword_case: "upper"` — keywords are uppercased (SELECT, FROM, WHERE); use "lower" or "preserve" to change
• Default `tab_width: 2` — indentation uses 2 spaces; tab_width: 4 widens; use_tabs: true switches to tab chars
• Default dialect is `"sql"` (generic ANSI) — dialect-specific syntax (MERGE for transactsql, NOW() for mysql) formats correctly when dialect is specified
• Returns structured JSON — {formatted: string, dialect: string, line_count: number}, not plain text
• First call ~175ms (JIT) — subsequent dialect-specific calls range 13-463ms, p50=138ms across 12 calls. Dialect changes are slower (~338-463ms for first postgresql/mysql call)
• Empty SQL returns `{formatted: "", line_count: 1}` — no error on empty input
• Handles complex SQL well — CTEs, window functions (ROW_NUMBER OVER PARTITION BY), MERGE statements, subqueries all format correctly with proper nesting
• `qty-1` formats as `qty -1` (adds space before minus) — cosmetic, not a bug