#Codegen

printSync generates source text from a program AST. It is exported by @swc-next/core, @swc-next/codegen, and the initialized @swc-next/core-browser package.

#Signature and inputs

declare function printSync(
  program: Program,
  options?: CodegenOptions | null,
  meta?: Partial<AstMeta>,
): CodegenOutput;

Input nodes follow the AST declarations shared with the parser.

#CodegenOptions

Omitted fields use their defaults. Non-boolean option values, such as minify: 1, throw a TypeError.

#asciiOnly

Type: boolean

Default: false

Use escape sequences for non-ASCII characters in identifiers, strings, and supported template text. With false, the printer can emit Unicode characters directly.

import { parseSync, printSync } from "@swc-next/core";

const { program } = parseSync('const greeting = "你好";');
printSync(program, { asciiOnly: false }).code;
// 'const greeting = "你好";\n'

printSync(program, { asciiOnly: true }).code;
// 'const greeting = "\\u4F60\\u597D";\n'

Escaping is context-dependent. Regular expression bodies and tagged template raw text are preserved, so this option does not guarantee that the entire output contains only ASCII. It does not change string values at runtime.

#inlineScript

Type: boolean

Default: false

Escape case-insensitive </script sequences in strings and untagged templates, and insert separators when neighboring tokens would form such a sequence. This helps prevent generated JavaScript text from closing an HTML <script> element early.

import { parseSync, printSync } from "@swc-next/core";

const { program } = parseSync('const text = "</script>";');
printSync(program, { inlineScript: false }).code;
// 'const text = "</script>";\n'

printSync(program, { inlineScript: true }).code;
// 'const text = "<\\/script>";\n'

The escape preserves the string's runtime value. Tagged template raw text is preserved because changing it would be observable by the tag function. This option therefore does not guarantee that arbitrary generated output can be embedded directly in HTML.

#minify

Type: boolean

Default: false

Use compact printing: remove optional whitespace, indentation, and line breaks, and choose shorter spellings where supported. Required separators remain so that adjacent tokens are parsed correctly. With false, print readable, indented code.

import { parseSync, printSync } from "@swc-next/core";

const { program } = parseSync("const answer = 1 + 2;");
printSync(program, { minify: false }).code;
// 'const answer = 1 + 2;\n'

printSync(program, { minify: true }).code;
// 'const answer=1+2;'

minify does not evaluate 1 + 2, remove unused declarations, or shorten variable names. Compression and name mangling belong to the separate Minifier component. Printing preserves TypeScript and JSX syntax with either setting.

#Metadata

interface AstMeta {
  readonly lang: "js" | "jsx" | "ts" | "tsx" | "dts";
  readonly sourceType: "module" | "script" | "commonjs" | "unambiguous";
}

Pass parseSync(...).meta to preserve the input language and source type. Explicit metadata takes precedence; when omitted, source type comes from the program and language is inferred from the tree. Printing does not remove TypeScript types or lower JSX.

import { parseSync, printSync } from "@swc-next/core";

const result = parseSync("export const answer: number = 42;", { lang: "ts" });
const { code } = printSync(result.program, { minify: true }, result.meta);

#CodegenOutput

interface CodegenOutput {
  code: string;
}

The JavaScript interface returns generated code only. It does not expose source maps or comment emission options; the parser's separate comment list is not passed to the printer.