diff --git a/pkg/sql/parser/ddl.go b/pkg/sql/parser/ddl.go
index 879070ed..b66d9ff3 100644
--- a/pkg/sql/parser/ddl.go
+++ b/pkg/sql/parser/ddl.go
@@ -95,6 +95,60 @@ func (p *Parser) parseCreateStatement() (ast.Statement, error) {
 		}
 		return stmt, nil
 	}
+
+	// Snowflake object-type extensions: STAGE, STREAM, TASK, PIPE, FILE FORMAT,
+	// WAREHOUSE, DATABASE, SCHEMA, ROLE, FUNCTION, PROCEDURE, SEQUENCE.
+	// Parse-only: record the object kind and name on a DescribeStatement
+	// placeholder, then consume the rest of the statement body permissively
+	// until ';' or EOF (tracking balanced parens so embedded expressions with
+	// semicolons inside string literals round-trip).
+	if p.dialect == string(keywords.DialectSnowflake) {
+		kind := strings.ToUpper(p.currentToken.Token.Value)
+		if kind == "FILE" && strings.EqualFold(p.peekToken().Token.Value, "FORMAT") {
+			p.advance() // FILE
+			kind = "FILE FORMAT"
+		}
+		switch kind {
+		case "STAGE", "STREAM", "TASK", "PIPE", "FILE FORMAT",
+			"WAREHOUSE", "DATABASE", "SCHEMA", "ROLE", "SEQUENCE",
+			"FUNCTION", "PROCEDURE":
+			p.advance() // Consume object-kind keyword
+			// Optional IF NOT EXISTS
+			if p.isType(models.TokenTypeIf) {
+				p.advance()
+				if p.isType(models.TokenTypeNot) {
+					p.advance()
+				}
+				if p.isType(models.TokenTypeExists) {
+					p.advance()
+				}
+			}
+			// Object name (qualified identifier)
+			name, _ := p.parseQualifiedName()
+			// Consume the rest of the statement body until ';' or EOF,
+			// tracking balanced parens.
+			depth := 0
+			for {
+				t := p.currentToken.Token.Type
+				if t == models.TokenTypeEOF {
+					break
+				}
+				if t == models.TokenTypeSemicolon && depth == 0 {
+					break
+				}
+				if t == models.TokenTypeLParen {
+					depth++
+				} else if t == models.TokenTypeRParen {
+					depth--
+				}
+				p.advance()
+			}
+			stub := ast.GetDescribeStatement()
+			stub.TableName = "CREATE " + kind + " " + name
+			return stub, nil
+		}
+	}
+
 	return nil, p.expectedError("TABLE, VIEW, MATERIALIZED VIEW, or INDEX after CREATE")
 }
 
diff --git a/pkg/sql/parser/snowflake_create_objects_test.go b/pkg/sql/parser/snowflake_create_objects_test.go
new file mode 100644
index 00000000..3ea5e874
--- /dev/null
+++ b/pkg/sql/parser/snowflake_create_objects_test.go
@@ -0,0 +1,47 @@
+// Copyright 2026 GoSQLX Authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+
+package parser_test
+
+import (
+	"testing"
+
+	"github.com/ajitpratap0/GoSQLX/pkg/gosqlx"
+	"github.com/ajitpratap0/GoSQLX/pkg/sql/keywords"
+)
+
+// TestSnowflakeCreateObjects verifies Snowflake CREATE statements for object
+// types beyond TABLE/VIEW/INDEX parse. These are currently consumed as
+// stub statements (body is not modeled on the AST). Regression for #483.
+func TestSnowflakeCreateObjects(t *testing.T) {
+	queries := map[string]string{
+		"create_stage": `CREATE STAGE my_stage URL='s3://bucket/path' CREDENTIALS=(AWS_KEY_ID='abc' AWS_SECRET_KEY='xyz')`,
+
+		"create_file_format": `CREATE FILE FORMAT my_csv TYPE = CSV FIELD_DELIMITER = ','`,
+
+		"create_stream": `CREATE STREAM my_stream ON TABLE events`,
+
+		"create_task": `CREATE TASK daily_refresh WAREHOUSE = compute_wh SCHEDULE = 'USING CRON 0 0 * * * UTC' AS INSERT INTO t SELECT 1`,
+
+		"create_or_replace_pipe": `CREATE OR REPLACE PIPE my_pipe AUTO_INGEST = TRUE AS COPY INTO t FROM @my_stage`,
+
+		"create_warehouse": `CREATE WAREHOUSE my_wh WITH WAREHOUSE_SIZE = 'SMALL'`,
+
+		"create_database": `CREATE DATABASE my_db`,
+
+		"create_schema_qualified": `CREATE SCHEMA analytics.my_schema`,
+
+		"create_role": `CREATE ROLE analyst`,
+
+		"create_if_not_exists_stage": `CREATE STAGE IF NOT EXISTS my_stage URL='s3://bucket'`,
+	}
+	for name, q := range queries {
+		q := q
+		t.Run(name, func(t *testing.T) {
+			if _, err := gosqlx.ParseWithDialect(q, keywords.DialectSnowflake); err != nil {
+				t.Fatalf("parse failed: %v", err)
+			}
+		})
+	}
+}