API Reference | Custom Tag Name Characters

Default syntax is $$tag(content)$$. Want @@tag(content)@@ or <<tag[content]>>? Change the syntax option.

Inline:   $$tag(content)$$
          ↑↑   ↑       ↑↑
      tagPrefix tagOpen endTag

          $$tag(arg | content)$$
                    ↑
               tagDivider

Raw:      $$tag(arg)%
                   ↑   raw content (no parsing)
              rawOpen
          %end$$
          ↑
          rawClose

Block:    $$tag(arg)*
                   ↑   block content (recursive parsing)
             blockOpen
          *end$$
          ↑
          blockClose

Escape:   \)  \\  \|
          ↑
          escapeChar

Want to customize syntax?
    │
    ├─ Most cases → createEasySyntax (recommended)
    │              Change base tokens, compounds auto-derive
    │
    └─ Full manual → createSyntax
                     Plain shallow merge, you manage everything

function createEasySyntax(
  overrides?: Partial<SyntaxInput> & { closeMiddle?: string }
): SyntaxConfig

You only change the easy-layer base inputs, compound tokens derive automatically:

escapeChar gotcha: The default is "\\" in JavaScript source — but that is a single \ character. The length of the string is what the user types in the DSL text. escapeChar: "\\" → user types \| to get a literal | (one keystroke). escapeChar: "~~" → user types ~~| to get a literal | (two keystrokes). If you set a multi-character escape, users must type the full sequence every time they escape.

Change prefix — compounds follow:

createEasySyntax({ tagPrefix: "@@" });
// endTag → ")@@"    rawClose → "%end@@"    blockClose → "*end@@"

Change prefix + closer:

createEasySyntax({ tagPrefix: "@@", tagClose: "]" });
// endTag → "]@@"    rawOpen → "]%"    blockOpen → "]*"

Change the shared close middle:

createEasySyntax({ tagPrefix: "@@", closeMiddle: "fin" });
// rawClose → "%fin@@"    blockClose → "*fin@@"

Just the divider:

createEasySyntax({ tagDivider: ";" });
// $$tag(arg1;arg2)$$

Explicit compound override wins over derivation:

createEasySyntax({
    tagPrefix: "@@",
    rawClose: "<<ENDRAW>>",  // explicit, derivation skipped
});
// endTag → ")@@" (derived)    rawClose → "<<ENDRAW>>" (your override)

function createSyntax(overrides?: Partial<SyntaxInput>): SyntaxConfig

Plain shallow merge with DEFAULT_SYNTAX. No auto-derivation, no validation. For fully irregular protocols.

const syntax = createSyntax({
    tagPrefix: "<<",
    tagOpen: "[",
    tagClose: "]",
    tagDivider: ";",
    endTag: "]>>",
    rawOpen: "]#",
    blockOpen: "]*",
    rawClose: "#end>>",
    blockClose: "*end>>",
    escapeChar: "~",
});
// <<bold[hello]>>   <<code[ts]#\nconst x = 1;\n#end>>

Warning: If you only override tagPrefix without updating endTag, inline tags won't close — endTag is still )$$ but tags open with the new prefix.

Hard constraint in the parser — break these and tags stop working:

Core rule: endTag, rawOpen, blockOpen must start with tagClose.

Why? The parser balances tagOpen/tagClose nesting in the argument section, lands on tagClose, then looks ahead to determine the tag form. If compound tokens don't start with tagClose, the lookahead never matches.

Using createEasySyntax guarantees these constraints automatically. closeMiddle only affects easy derivation; it does not relax these parser requirements.

DEFAULT_SYNTAX.tagPrefix   === "$$"
DEFAULT_SYNTAX.tagOpen     === "("
DEFAULT_SYNTAX.tagClose    === ")"
DEFAULT_SYNTAX.tagDivider  === "|"
DEFAULT_SYNTAX.endTag      === ")$$"
DEFAULT_SYNTAX.rawOpen     === ")%"
DEFAULT_SYNTAX.blockOpen   === ")*"
DEFAULT_SYNTAX.rawClose    === "%end$$"
DEFAULT_SYNTAX.blockClose  === "*end$$"
DEFAULT_SYNTAX.escapeChar  === "\\"