diff --git a/compiler/rustc_parse/messages.ftl b/compiler/rustc_parse/messages.ftl
index 747895c80469..1331d99c01ea 100644
--- a/compiler/rustc_parse/messages.ftl
+++ b/compiler/rustc_parse/messages.ftl
@@ -967,6 +967,7 @@ parse_unknown_start_of_token = unknown start of token: {$escaped}
     .sugg_quotes = Unicode characters '“' (Left Double Quotation Mark) and '”' (Right Double Quotation Mark) look like '{$ascii_str}' ({$ascii_name}), but are not
     .sugg_other = Unicode character '{$ch}' ({$u_name}) looks like '{$ascii_str}' ({$ascii_name}), but it is not
     .help_null = source files must contain UTF-8 encoded text, unexpected null bytes might occur when a different encoding is used
+    .help_invisible_char = invisible characters like '{$escaped}' are not usually visible in text editors
     .note_repeats = character appears {$repeats ->
         [one] once more
         *[other] {$repeats} more times
diff --git a/compiler/rustc_parse/src/errors.rs b/compiler/rustc_parse/src/errors.rs
index 3b72c9802afd..60e4a240c85e 100644
--- a/compiler/rustc_parse/src/errors.rs
+++ b/compiler/rustc_parse/src/errors.rs
@@ -2369,6 +2369,8 @@ pub(crate) struct UnknownTokenStart {
     pub null: Option<UnknownTokenNull>,
     #[subdiagnostic]
     pub repeat: Option<UnknownTokenRepeat>,
+    #[subdiagnostic]
+    pub invisible: Option<InvisibleCharacter>,
 }
 
 #[derive(Subdiagnostic)]
@@ -2409,6 +2411,10 @@ pub(crate) struct UnknownTokenRepeat {
     pub repeats: usize,
 }
 
+#[derive(Subdiagnostic)]
+#[help(parse_help_invisible_char)]
+pub(crate) struct InvisibleCharacter;
+
 #[derive(Subdiagnostic)]
 #[help(parse_help_null)]
 pub(crate) struct UnknownTokenNull;
diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs
index 94ae35c19582..7c969dd7f9f4 100644
--- a/compiler/rustc_parse/src/lexer/mod.rs
+++ b/compiler/rustc_parse/src/lexer/mod.rs
@@ -36,6 +36,10 @@ use unescape_error_reporting::{emit_unescape_error, escaped_char};
 #[cfg(target_pointer_width = "64")]
 rustc_data_structures::static_assert_size!(rustc_lexer::Token, 12);
 
+const INVISIBLE_CHARACTERS: [char; 8] = [
+    '\u{200b}', '\u{200c}', '\u{2060}', '\u{2061}', '\u{2062}', '\u{00ad}', '\u{034f}', '\u{061c}',
+];
+
 #[derive(Clone, Debug)]
 pub(crate) struct UnmatchedDelim {
     pub found_delim: Option<Delimiter>,
@@ -456,6 +460,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                         escaped: escaped_char(c),
                         sugg,
                         null: if c == '\x00' { Some(errors::UnknownTokenNull) } else { None },
+                        invisible: if INVISIBLE_CHARACTERS.contains(&c) { Some(errors::InvisibleCharacter) } else { None },
                         repeat: if repeats > 0 {
                             swallow_next_invalid = repeats;
                             Some(errors::UnknownTokenRepeat { repeats })
diff --git a/tests/ui/lexer/lex-invisible-characters.rs b/tests/ui/lexer/lex-invisible-characters.rs
new file mode 100644
index 000000000000..2db72b8475dc
--- /dev/null
+++ b/tests/ui/lexer/lex-invisible-characters.rs
@@ -0,0 +1,6 @@
+// Provide extra help when a user has an invisible character in their code
+
+fn main​() {
+    //~^ ERROR unknown start of token: \u{200b}
+    //~| HELP invisible characters like '\u{200b}' are not usually visible in text editors
+}
diff --git a/tests/ui/lexer/lex-invisible-characters.stderr b/tests/ui/lexer/lex-invisible-characters.stderr
new file mode 100644
index 000000000000..ddac0f4e9325
--- /dev/null
+++ b/tests/ui/lexer/lex-invisible-characters.stderr
@@ -0,0 +1,10 @@
+error: unknown start of token: \u{200b}
+  --> $DIR/lex-invisible-characters.rs:3:8
+   |
+LL | fn main​() {
+   |        ^
+   |
+   = help: invisible characters like '\u{200b}' are not usually visible in text editors
+
+error: aborting due to 1 previous error
+