diff --git a/CHANGELOG.md b/CHANGELOG.md
index 7bb44e8d..536031b7 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -14,6 +14,8 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
     middle are adjusted to retain their original significant leading whitespace.
   2. Internal line endings are normalised to match the configured line ending.
   3. Invalid multiline strings are left untouched.
+- All keywords are now set to lowercase. This includes reserved words (like `begin` and `const`) but also includes
+  non-reserved words (like `absolute` and `override`) when used in the context that makes them special.
 
 ## [0.6.0] - 2025-08-25
 
diff --git a/core/CHANGELOG.md b/core/CHANGELOG.md
index 206df1e3..35147f9c 100644
--- a/core/CHANGELOG.md
+++ b/core/CHANGELOG.md
@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 
 - Added setting to enable internal formatting of multiline strings: `OptimisingLineFormatterSettings.format_multiline_strings`.
+- Added `LowercaseKeywords` to convert instances of keywords to lowercase.
 
 ### Changed
 
diff --git a/core/src/rules/lowercase_keywords.rs b/core/src/rules/lowercase_keywords.rs
new file mode 100644
index 00000000..b1e11d15
--- /dev/null
+++ b/core/src/rules/lowercase_keywords.rs
@@ -0,0 +1,59 @@
+use crate::prelude::*;
+
+pub struct LowercaseKeywords {}
+
+impl LogicalLineFileFormatter for LowercaseKeywords {
+    fn format(&self, formatted_tokens: &mut FormattedTokens<'_>, _input: &[LogicalLine]) {
+        for (tok, _) in formatted_tokens.tokens_mut() {
+            let Ok(tok) = tok else { continue };
+            if matches!(tok.get_token_type(), TokenType::Keyword(_))
+                && tok.get_content().bytes().any(|b| b.is_ascii_uppercase())
+            {
+                tok.set_content(tok.get_content().to_ascii_lowercase());
+            }
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    fn formatter() -> Formatter {
+        Formatter::builder()
+            .lexer(DelphiLexer {})
+            .parser(DelphiLogicalLineParser {})
+            .token_ignorer(FormattingToggler {})
+            .file_formatter(LowercaseKeywords {})
+            .reconstructor(default_test_reconstructor())
+            .build()
+    }
+
+    formatter_test_group!(
+        tests,
+        full_uppercase = {
+            "BEGIN END",
+            "begin end"
+        },
+        partial_uppercase = {
+            "begIn enD",
+            "begin end"
+        },
+        lowercase = {
+            "begin end",
+            "begin end"
+        },
+        impure_keyword_is_ignored = {
+            "ABSOLUTE := 0",
+            "ABSOLUTE := 0",
+        },
+        impure_keyword_is_formatted = {
+            "var a: b ABSOLUTE c",
+            "var a: b absolute c",
+        },
+        ignored_tokens = {
+            "{pasfmt off} BEGIN {pasfmt on} END",
+            "{pasfmt off} BEGIN {pasfmt on} end",
+        }
+    );
+}
diff --git a/core/src/rules/mod.rs b/core/src/rules/mod.rs
index 37561412..1472765a 100644
--- a/core/src/rules/mod.rs
+++ b/core/src/rules/mod.rs
@@ -4,6 +4,7 @@ pub mod eof_newline;
 pub mod formatting_toggle;
 pub mod generics_consolidator;
 pub mod ignore_asm_instructions;
+pub mod lowercase_keywords;
 pub mod optimising_line_formatter;
 pub mod token_spacing;
 
@@ -13,5 +14,6 @@ pub use eof_newline::*;
 pub use formatting_toggle::*;
 pub use generics_consolidator::*;
 pub use ignore_asm_instructions::*;
+pub use lowercase_keywords::*;
 pub use optimising_line_formatter::*;
 pub use token_spacing::*;
diff --git a/front-end/src/lib.rs b/front-end/src/lib.rs
index 93d68319..bed2d258 100644
--- a/front-end/src/lib.rs
+++ b/front-end/src/lib.rs
@@ -322,6 +322,7 @@ pub fn make_formatter(config: &FormattingConfig) -> Formatter {
         .token_ignorer(FormattingToggler {})
         .token_ignorer(IgnoreAsmIstructions {})
         .file_formatter(TokenSpacing {})
+        .file_formatter(LowercaseKeywords {})
         .line_formatter(FormatterSelector::new(
             |logical_line_type| match logical_line_type {
                 LogicalLineType::Eof => Some(eof_newline_formatter),