diff --git a/src/parse/mod.rs b/src/parse/mod.rs
index bdecc31f8..1d893ad9d 100644
--- a/src/parse/mod.rs
+++ b/src/parse/mod.rs
@@ -102,14 +102,12 @@ fn node(p: &mut Parser, at_start: &mut bool) -> Option<Node> {
         Token::Star => Node::Strong(span),
         Token::Underscore => Node::Emph(span),
         Token::Raw(t) => raw(p, t),
-        Token::Hashtag if *at_start => return Some(heading(p)),
+        Token::Eq if *at_start => return Some(heading(p)),
         Token::Hyph if *at_start => return Some(list_item(p)),
         Token::Numbering(number) if *at_start => return Some(enum_item(p, number)),
 
         // Line-based markup that is not currently at the start of the line.
-        Token::Hashtag | Token::Hyph | Token::Numbering(_) => {
-            Node::Text(p.peek_src().into())
-        }
+        Token::Eq | Token::Hyph | Token::Numbering(_) => Node::Text(p.peek_src().into()),
 
         // Hashtag + keyword / identifier.
         Token::Ident(_)
@@ -183,11 +181,11 @@ fn raw(p: &mut Parser, token: RawToken) -> Node {
 /// Parse a heading.
 fn heading(p: &mut Parser) -> Node {
     let start = p.next_start();
-    p.assert(Token::Hashtag);
+    p.assert(Token::Eq);
 
     // Count depth.
     let mut level: usize = 1;
-    while p.eat_if(Token::Hashtag) {
+    while p.eat_if(Token::Eq) {
         level += 1;
     }
 
diff --git a/src/parse/scanner.rs b/src/parse/scanner.rs
index fad44e89f..af88aa684 100644
--- a/src/parse/scanner.rs
+++ b/src/parse/scanner.rs
@@ -88,12 +88,12 @@ impl<'s> Scanner<'s> {
 
     /// Checks whether the next char fulfills a condition.
     ///
-    /// Returns `false` if there is no next char.
-    pub fn check<F>(&self, f: F) -> bool
+    /// Returns `default` if there is no next char.
+    pub fn check_or<F>(&self, default: bool, f: F) -> bool
     where
         F: FnOnce(char) -> bool,
     {
-        self.peek().map(f).unwrap_or(false)
+        self.peek().map_or(default, f)
     }
 
     /// The previous index in the source string.
diff --git a/src/parse/tokens.rs b/src/parse/tokens.rs
index f21967402..aebe7b70d 100644
--- a/src/parse/tokens.rs
+++ b/src/parse/tokens.rs
@@ -89,19 +89,22 @@ impl<'s> Iterator for Tokens<'s> {
 impl<'s> Tokens<'s> {
     fn markup(&mut self, start: usize, c: char) -> Token<'s> {
         match c {
+            // Escape sequences.
+            '\\' => self.backslash(),
+
+            // Keywords and identifiers.
+            '#' => self.hash(),
+
             // Markup.
             '~' => Token::Tilde,
             '*' => Token::Star,
             '_' => Token::Underscore,
-            '\\' => self.backslash(),
             '`' => self.raw(),
             '$' => self.math(),
             '-' => self.hyph(start),
+            '=' if self.s.check_or(true, |c| c == '=' || c.is_whitespace()) => Token::Eq,
             c if c == '.' || c.is_ascii_digit() => self.numbering(start, c),
 
-            // Headings, keywords and identifiers.
-            '#' => self.hash(start),
-
             // Plain text.
             _ => self.text(start),
         }
@@ -143,7 +146,7 @@ impl<'s> Tokens<'s> {
 
             // Numbers.
             c if c.is_ascii_digit()
-                || (c == '.' && self.s.check(|n| n.is_ascii_digit())) =>
+                || (c == '.' && self.s.check_or(false, |n| n.is_ascii_digit())) =>
             {
                 self.number(start, c)
             }
@@ -157,7 +160,7 @@ impl<'s> Tokens<'s> {
 
     fn whitespace(&mut self, first: char) -> Token<'s> {
         // Fast path for just a single space
-        if first == ' ' && !self.s.check(char::is_whitespace) {
+        if first == ' ' && self.s.check_or(true, |c| !c.is_whitespace()) {
             Token::Space(0)
         } else {
             self.s.uneat();
@@ -188,8 +191,10 @@ impl<'s> Tokens<'s> {
                 '/' => true,
                 // Parentheses.
                 '[' | ']' | '{' | '}' => true,
+                // Code.
+                '#' => true,
                 // Markup.
-                '#' | '~' | '*' | '_' | '`' | '$' | '-' => true,
+                '~' | '*' | '_' | '`' | '$' | '-' => true,
                 // Escaping.
                 '\\' => true,
                 // Just text.
@@ -233,18 +238,16 @@ impl<'s> Tokens<'s> {
         }
     }
 
-    fn hash(&mut self, start: usize) -> Token<'s> {
-        if self.s.check(is_id_start) {
+    fn hash(&mut self) -> Token<'s> {
+        if self.s.check_or(false, is_id_start) {
             let read = self.s.eat_while(is_id_continue);
             if let Some(keyword) = keyword(read) {
                 keyword
             } else {
                 Token::Ident(read)
             }
-        } else if self.s.check(|c| c != '#' && !c.is_whitespace()) {
-            Token::Text(self.s.eaten_from(start))
         } else {
-            Token::Hashtag
+            Token::Invalid("#")
         }
     }
 
@@ -255,10 +258,10 @@ impl<'s> Tokens<'s> {
             } else {
                 Token::HyphHyph
             }
-        } else if self.s.check(|c| !c.is_whitespace()) {
-            Token::Text(self.s.eaten_from(start))
-        } else {
+        } else if self.s.check_or(true, char::is_whitespace) {
             Token::Hyph
+        } else {
+            Token::Text(self.s.eaten_from(start))
         }
     }
 
@@ -274,11 +277,11 @@ impl<'s> Tokens<'s> {
             None
         };
 
-        if self.s.check(|c| !c.is_whitespace()) {
-            return Token::Text(self.s.eaten_from(start));
+        if self.s.check_or(true, char::is_whitespace) {
+            Token::Numbering(number)
+        } else {
+            Token::Text(self.s.eaten_from(start))
         }
-
-        Token::Numbering(number)
     }
 
     fn raw(&mut self) -> Token<'s> {
@@ -663,8 +666,8 @@ mod tests {
         // Test code symbols in text.
         t!(Markup[" /"]: "a():\"b" => Text("a():\"b"));
         t!(Markup[" /"]: ";:,|/+"  => Text(";:,|"), Text("/+"));
-        t!(Markup[" /"]: "#-a"     => Text("#"), Text("-"), Text("a"));
-        t!(Markup[" "]: "#123"     => Text("#"), Text("123"));
+        t!(Markup[" /"]: "=-a"     => Text("="), Text("-"), Text("a"));
+        t!(Markup[" "]: "#123"     => Invalid("#"), Text("123"));
 
         // Test text ends.
         t!(Markup[""]: "hello " => Text("hello"), Space(0));
@@ -712,8 +715,8 @@ mod tests {
         // Test markup tokens.
         t!(Markup[" a1"]: "*"   => Star);
         t!(Markup: "_"          => Underscore);
-        t!(Markup[""]: "###"    => Hashtag, Hashtag, Hashtag);
-        t!(Markup["a1/"]: "# "  => Hashtag, Space(0));
+        t!(Markup[""]: "==="    => Eq, Eq, Eq);
+        t!(Markup["a1/"]: "= "  => Eq, Space(0));
         t!(Markup: "~"          => Tilde);
         t!(Markup[" "]: r"\"    => Backslash);
         t!(Markup["a "]: r"a--" => Text("a"), HyphHyph);
@@ -776,7 +779,7 @@ mod tests {
         for &(s, t) in &list {
             t!(Markup[" "]: format!("#{}", s) => t);
             t!(Markup[" "]: format!("#{0}#{0}", s) => t, t);
-            t!(Markup[" /"]: format!("# {}", s) => Token::Hashtag, Space(0), Text(s));
+            t!(Markup[" /"]: format!("# {}", s) => Token::Invalid("#"), Space(0), Text(s));
         }
 
         for &(s, t) in &list {
diff --git a/src/pretty.rs b/src/pretty.rs
index 17609ceaf..ff8841f3e 100644
--- a/src/pretty.rs
+++ b/src/pretty.rs
@@ -169,7 +169,7 @@ impl Pretty for RawNode {
 impl Pretty for HeadingNode {
     fn pretty(&self, p: &mut Printer) {
         for _ in 0 .. self.level {
-            p.push('#');
+            p.push('=');
         }
         p.push(' ');
         self.body.pretty(p);
@@ -653,7 +653,7 @@ mod tests {
         roundtrip("\\ ");
         roundtrip("\n\n");
         roundtrip("hi");
-        roundtrip("# *Ok*");
+        roundtrip("= *Ok*");
         roundtrip("- Ok");
 
         // Raw.
diff --git a/src/syntax/token.rs b/src/syntax/token.rs
index 32a14d58f..425dac108 100644
--- a/src/syntax/token.rs
+++ b/src/syntax/token.rs
@@ -19,8 +19,6 @@ pub enum Token<'s> {
     Star,
     /// An underscore: `_`.
     Underscore,
-    /// A single hashtag: `#`.
-    Hashtag,
     /// A tilde: `~`.
     Tilde,
     /// Two hyphens: `--`.
@@ -215,7 +213,6 @@ impl<'s> Token<'s> {
             Self::RightParen => "closing paren",
             Self::Star => "star",
             Self::Underscore => "underscore",
-            Self::Hashtag => "hashtag",
             Self::Tilde => "tilde",
             Self::HyphHyph => "en dash",
             Self::HyphHyphHyph => "em dash",
diff --git a/tests/ref/code/let.png b/tests/ref/code/let.png
index ae360b6b6..e9c000644 100644
Binary files a/tests/ref/code/let.png and b/tests/ref/code/let.png differ
diff --git a/tests/ref/markup/heading.png b/tests/ref/markup/heading.png
index f69bb887d..35a589bd8 100644
Binary files a/tests/ref/markup/heading.png and b/tests/ref/markup/heading.png differ
diff --git a/tests/typ/code/call-wide.typ b/tests/typ/code/call-wide.typ
index 54da540bd..996a052d1 100644
--- a/tests/typ/code/call-wide.typ
+++ b/tests/typ/code/call-wide.typ
@@ -7,7 +7,7 @@
 
 ---
 // Test in heading.
-# A #align!(right) B
+= A #align!(right) B
 C
 
 ---
diff --git a/tests/typ/code/importable/chap1.typ b/tests/typ/code/importable/chap1.typ
index a0f38d0d6..06a4c1a1a 100644
--- a/tests/typ/code/importable/chap1.typ
+++ b/tests/typ/code/importable/chap1.typ
@@ -2,7 +2,7 @@
 
 #let name = "Klaus"
 
-## Chapter 1
+== Chapter 1
 #name stood in a field of wheat. There was nothing of particular interest about
 the field #name just casually surveyed for any paths on which the corn would not
 totally ruin his semi-new outdorsy jacket but then again, most of us spend
diff --git a/tests/typ/code/importable/chap2.typ b/tests/typ/code/importable/chap2.typ
index 51e116aa8..d4aedc60d 100644
--- a/tests/typ/code/importable/chap2.typ
+++ b/tests/typ/code/importable/chap2.typ
@@ -2,7 +2,7 @@
 
 #let name = "Klaus"
 
-## Chapter 2
+== Chapter 2
 Their motivations, however, were pretty descript, so to speak. #name had not yet
 conceptualized their consequences, but that should change pretty quickly. #name
 approached the center of the field and picked up a 4-foot long disk made from
diff --git a/tests/typ/code/include.typ b/tests/typ/code/include.typ
index 8080c6e85..166c3945c 100644
--- a/tests/typ/code/include.typ
+++ b/tests/typ/code/include.typ
@@ -1,7 +1,7 @@
 // Test include statements.
 
 ---
-# Document
+= Document
 
 // Include a file
 #include "importable/chap1.typ"
diff --git a/tests/typ/code/let.typ b/tests/typ/code/let.typ
index 1a0bb10ea..d788a5639 100644
--- a/tests/typ/code/let.typ
+++ b/tests/typ/code/let.typ
@@ -79,6 +79,6 @@ Three
 // Error: 9 expected expression
 #let v =
 
-// Should output `= 1`.
+// Should output a heading `1`.
 // Error: 6-9 expected identifier, found string
 #let "v" = 1
diff --git a/tests/typ/coma.typ b/tests/typ/coma.typ
index 5074bd9ea..161ac832d 100644
--- a/tests/typ/coma.typ
+++ b/tests/typ/coma.typ
@@ -30,7 +30,7 @@
 // the parentheses.
 #align(center)[
     // Markdown-like syntax for headings.
-    #### 3. Übungsblatt Computerorientierte Mathematik II #v(4mm)
+    ==== 3. Übungsblatt Computerorientierte Mathematik II #v(4mm)
     *Abgabe: 03.05.2019* (bis 10:10 Uhr in MA 001) #v(4mm)
     *Alle Antworten sind zu beweisen.*
 ]
diff --git a/tests/typ/markup/heading.typ b/tests/typ/markup/heading.typ
index 8787d959a..e0dbd999a 100644
--- a/tests/typ/markup/heading.typ
+++ b/tests/typ/markup/heading.typ
@@ -4,39 +4,39 @@
 // Different number of hashtags.
 
 // Valid levels.
-# Level 1
-### Level 2
-###### Level 6
+= Level 1
+=== Level 2
+====== Level 6
 
 // Too many hashtags.
 // Warning: 1-8 should not exceed depth 6
-####### Level 7
+======= Level 7
 
 ---
 // Heading vs. no heading.
 
 // Parsed as headings if at start of the context.
-/**/ # Level 1
-{[## Level 2]}
-#box[### Level 3]
+/**/ = Level 1
+{[== Level 2]}
+#box[=== Level 3]
 
 // Not at the start of the context.
-No # heading
+No = heading
 
 // Escaped.
-\# No heading
+\= No heading
 
 ---
 // While indented at least as much as the start, the heading continues.
 
-# This
+= This
   is
     indented.
 
-#  This
+=  This
   is not.
 
 // Code blocks continue heading.
-# A {
+= A {
     "B"
 }
diff --git a/tools/support/typst.tmLanguage.json b/tools/support/typst.tmLanguage.json
index e1f416535..33e4c3db7 100644
--- a/tools/support/typst.tmLanguage.json
+++ b/tools/support/typst.tmLanguage.json
@@ -75,7 +75,7 @@
         {
           "name": "markup.heading.typst",
           "contentName": "entity.name.section.typst",
-          "begin": "^\\s*#{1,6}\\s+",
+          "begin": "^\\s*={1,6}\\s+",
           "end": "\n",
           "beginCaptures": { "0": { "name": "punctuation.definition.heading.typst" } },
           "patterns": [{ "include": "#markup" }]