diff --git a/services/web/frontend/js/features/source-editor/lezer-typst/typst.grammar b/services/web/frontend/js/features/source-editor/lezer-typst/typst.grammar
index 4b0b52d027..ad485b1704 100644
--- a/services/web/frontend/js/features/source-editor/lezer-typst/typst.grammar
+++ b/services/web/frontend/js/features/source-editor/lezer-typst/typst.grammar
@@ -31,12 +31,15 @@ item {
 // HeadingMark is produced by an external tokenizer that enforces the
 // start-of-line constraint and captures the "=+" prefix + trailing space.
 Heading { HeadingMark HeadingTitle }
-// headingTitleItem+ (not *) so HeadingTitle is never empty: the generator
-// would otherwise produce a shift/reduce conflict on "*" after HeadingMark,
-// unable to tell if "*" starts a Strong inside the heading or at doc level.
+// Strong and Emphasis are intentionally excluded from headingTitleItem.
+// Including them causes an LALR(1) conflict: since Strong/Emphasis also appear
+// in document-level `item`, the LR automaton merges heading-title states with
+// document-item states, making "*" ambiguous (Strong opener vs. end of heading).
+// Instead, HeadingText is widened to consume "*" and "_" as plain text inside
+// headings — they are not interpreted as markup delimiters there.
 HeadingTitle { headingTitleItem+ }
 headingTitleItem {
-  Strong | Emphasis | CodeExpr | InlineMath | RawInline | Label | Ref | HeadingText
+  CodeExpr | InlineMath | RawInline | Label | Ref | HeadingText
 }
 
 // ── Comments ──────────────────────────────────────────────────────────────
@@ -186,7 +189,7 @@ Escape { "\\" EscapeChar }
   MathContent { ![$\n]+ }
 
   // Text tokens for different markup contexts; each excludes its own delimiters.
-  HeadingText  { ![\n*_$#`<@\\]+  }
+  HeadingText  { ![\n$#`<@\\]+  }
   StrongText   { ![\n*$#`@\\]+    }
   EmphText     { ![\n_$#`@\\]+    }