diff --git a/Cargo.toml b/Cargo.toml
index 5e707ae79..1e6ec664d 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -14,7 +14,7 @@ finite automata and guarantees linear time matching on all inputs.
 categories = ["text-processing"]
 autotests = false
 exclude = ["/scripts/*", "/.github/*"]
-edition = "2018"
+edition = "2021"
 rust-version = "1.60.0"
 
 [workspace]
diff --git a/regex-syntax/Cargo.toml b/regex-syntax/Cargo.toml
index 6c92717e3..e30de004b 100644
--- a/regex-syntax/Cargo.toml
+++ b/regex-syntax/Cargo.toml
@@ -14,7 +14,8 @@ rust-version = "1.60.0"
 # Features are documented in the "Crate features" section of the crate docs:
 # https://docs.rs/regex-syntax/*/#crate-features
 [features]
-default = ["unicode"]
+default = ["std", "unicode"]
+std = []
 
 unicode = [
   "unicode-age",
diff --git a/regex-syntax/src/ast/mod.rs b/regex-syntax/src/ast/mod.rs
index 7c9dae7a0..7329fabbe 100644
--- a/regex-syntax/src/ast/mod.rs
+++ b/regex-syntax/src/ast/mod.rs
@@ -2,8 +2,9 @@
 Defines an abstract syntax for regular expressions.
 */
 
-use std::cmp::Ordering;
-use std::fmt;
+use core::cmp::Ordering;
+
+use alloc::{boxed::Box, string::String, vec, vec::Vec};
 
 pub use crate::ast::visitor::{visit, Visitor};
 
@@ -174,23 +175,24 @@ pub enum ErrorKind {
     UnsupportedLookAround,
 }
 
+#[cfg(feature = "std")]
 impl std::error::Error for Error {}
 
-impl fmt::Display for Error {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for Error {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         crate::error::Formatter::from(self).fmt(f)
     }
 }
 
-impl fmt::Display for ErrorKind {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for ErrorKind {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         use self::ErrorKind::*;
         match *self {
             CaptureLimitExceeded => write!(
                 f,
                 "exceeded the maximum number of \
                  capturing groups ({})",
-                ::std::u32::MAX
+                u32::MAX
             ),
             ClassEscapeInvalid => {
                 write!(f, "invalid escape sequence found in character class")
@@ -283,8 +285,8 @@ pub struct Span {
     pub end: Position,
 }
 
-impl fmt::Debug for Span {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Debug for Span {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         write!(f, "Span({:?}, {:?})", self.start, self.end)
     }
 }
@@ -316,8 +318,8 @@ pub struct Position {
     pub column: usize,
 }
 
-impl fmt::Debug for Position {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Debug for Position {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         write!(
             f,
             "Position(o: {:?}, l: {:?}, c: {:?})",
@@ -497,8 +499,8 @@ impl Ast {
 ///
 /// This implementation uses constant stack space and heap space proportional
 /// to the size of the `Ast`.
-impl fmt::Display for Ast {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for Ast {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         use crate::ast::print::Printer;
         Printer::new().print(self, f)
     }
@@ -1315,7 +1317,7 @@ pub enum Flag {
 /// space but heap space proportional to the depth of the `Ast`.
 impl Drop for Ast {
     fn drop(&mut self) {
-        use std::mem;
+        use core::mem;
 
         match *self {
             Ast::Empty(_)
@@ -1365,7 +1367,7 @@ impl Drop for Ast {
 /// stack space but heap space proportional to the depth of the `ClassSet`.
 impl Drop for ClassSet {
     fn drop(&mut self) {
-        use std::mem;
+        use core::mem;
 
         match *self {
             ClassSet::Item(ref item) => match *item {
diff --git a/regex-syntax/src/ast/parse.rs b/regex-syntax/src/ast/parse.rs
index 6e9c9aca0..f730ee659 100644
--- a/regex-syntax/src/ast/parse.rs
+++ b/regex-syntax/src/ast/parse.rs
@@ -2,17 +2,26 @@
 This module provides a regular expression parser.
 */
 
-use std::borrow::Borrow;
-use std::cell::{Cell, RefCell};
-use std::mem;
-use std::result;
-
-use crate::ast::{self, Ast, Position, Span};
-use crate::either::Either;
-
-use crate::is_meta_character;
-
-type Result<T> = result::Result<T, ast::Error>;
+use core::{
+    borrow::Borrow,
+    cell::{Cell, RefCell},
+    mem,
+};
+
+use alloc::{
+    boxed::Box,
+    string::{String, ToString},
+    vec,
+    vec::Vec,
+};
+
+use crate::{
+    ast::{self, Ast, Position, Span},
+    either::Either,
+    is_meta_character,
+};
+
+type Result<T> = core::result::Result<T, ast::Error>;
 
 /// A primitive is an expression with no sub-expressions. This includes
 /// literals, assertions and non-set character classes. This representation
@@ -1533,9 +1542,6 @@ impl<'s, P: Borrow<Parser>> ParserI<'s, P> {
     /// Assuming the preconditions are met, this routine can never fail.
     #[inline(never)]
     fn parse_octal(&self) -> ast::Literal {
-        use std::char;
-        use std::u32;
-
         assert!(self.parser().octal);
         assert!('0' <= self.char() && self.char() <= '7');
         let start = self.pos();
@@ -1600,9 +1606,6 @@ impl<'s, P: Borrow<Parser>> ParserI<'s, P> {
         &self,
         kind: ast::HexLiteralKind,
     ) -> Result<ast::Literal> {
-        use std::char;
-        use std::u32;
-
         let mut scratch = self.parser().scratch.borrow_mut();
         scratch.clear();
 
@@ -1646,9 +1649,6 @@ impl<'s, P: Borrow<Parser>> ParserI<'s, P> {
         &self,
         kind: ast::HexLiteralKind,
     ) -> Result<ast::Literal> {
-        use std::char;
-        use std::u32;
-
         let mut scratch = self.parser().scratch.borrow_mut();
         scratch.clear();
 
@@ -2146,7 +2146,7 @@ impl<'p, 's, P: Borrow<Parser>> NestLimiter<'p, 's, P> {
         let new = self.depth.checked_add(1).ok_or_else(|| {
             self.p.error(
                 span.clone(),
-                ast::ErrorKind::NestLimitExceeded(::std::u32::MAX),
+                ast::ErrorKind::NestLimitExceeded(u32::MAX),
             )
         })?;
         let limit = self.p.parser().nest_limit;
@@ -2297,11 +2297,14 @@ fn specialize_err<T>(
 
 #[cfg(test)]
 mod tests {
-    use std::ops::Range;
+    use core::ops::Range;
+
+    use alloc::format;
 
-    use super::{Parser, ParserBuilder, ParserI, Primitive};
     use crate::ast::{self, Ast, Position, Span};
 
+    use super::*;
+
     // Our own assert_eq, which has slightly better formatting (but honestly
     // still kind of crappy).
     macro_rules! assert_eq {
@@ -4272,7 +4275,7 @@ bar
                 Ok(Primitive::Literal(ast::Literal {
                     span: span(0..pat.len()),
                     kind: ast::LiteralKind::Octal,
-                    c: ::std::char::from_u32(i).unwrap(),
+                    c: char::from_u32(i).unwrap(),
                 }))
             );
         }
@@ -4347,7 +4350,7 @@ bar
                 Ok(Primitive::Literal(ast::Literal {
                     span: span(0..pat.len()),
                     kind: ast::LiteralKind::HexFixed(ast::HexLiteralKind::X),
-                    c: ::std::char::from_u32(i).unwrap(),
+                    c: char::from_u32(i).unwrap(),
                 }))
             );
         }
@@ -4378,7 +4381,7 @@ bar
     #[test]
     fn parse_hex_four() {
         for i in 0..65536 {
-            let c = match ::std::char::from_u32(i) {
+            let c = match char::from_u32(i) {
                 None => continue,
                 Some(c) => c,
             };
@@ -4442,7 +4445,7 @@ bar
     #[test]
     fn parse_hex_eight() {
         for i in 0..65536 {
-            let c = match ::std::char::from_u32(i) {
+            let c = match char::from_u32(i) {
                 None => continue,
                 Some(c) => c,
             };
diff --git a/regex-syntax/src/ast/print.rs b/regex-syntax/src/ast/print.rs
index f6b2462c0..e6c000d57 100644
--- a/regex-syntax/src/ast/print.rs
+++ b/regex-syntax/src/ast/print.rs
@@ -2,10 +2,13 @@
 This module provides a regular expression printer for `Ast`.
 */
 
-use std::fmt;
+use core::fmt;
 
-use crate::ast::visitor::{self, Visitor};
-use crate::ast::{self, Ast};
+use crate::ast::{
+    self,
+    visitor::{self, Visitor},
+    Ast,
+};
 
 /// A builder for constructing a printer.
 ///
@@ -395,9 +398,12 @@ impl<W: fmt::Write> Writer<W> {
 
 #[cfg(test)]
 mod tests {
-    use super::Printer;
+    use alloc::string::String;
+
     use crate::ast::parse::ParserBuilder;
 
+    use super::*;
+
     fn roundtrip(given: &str) {
         roundtrip_with(|b| b, given);
     }
diff --git a/regex-syntax/src/ast/visitor.rs b/regex-syntax/src/ast/visitor.rs
index 78ee487cf..03f8bf963 100644
--- a/regex-syntax/src/ast/visitor.rs
+++ b/regex-syntax/src/ast/visitor.rs
@@ -1,4 +1,4 @@
-use std::fmt;
+use alloc::{vec, vec::Vec};
 
 use crate::ast::{self, Ast};
 
@@ -475,8 +475,8 @@ impl<'a> ClassInduct<'a> {
     }
 }
 
-impl<'a> fmt::Debug for ClassFrame<'a> {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl<'a> core::fmt::Debug for ClassFrame<'a> {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         let x = match *self {
             ClassFrame::Union { .. } => "Union",
             ClassFrame::Binary { .. } => "Binary",
@@ -487,8 +487,8 @@ impl<'a> fmt::Debug for ClassFrame<'a> {
     }
 }
 
-impl<'a> fmt::Debug for ClassInduct<'a> {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl<'a> core::fmt::Debug for ClassInduct<'a> {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         let x = match *self {
             ClassInduct::Item(it) => match *it {
                 ast::ClassSetItem::Empty(_) => "Item(Empty)",
diff --git a/regex-syntax/src/error.rs b/regex-syntax/src/error.rs
index 6e7fa7466..a10230a87 100644
--- a/regex-syntax/src/error.rs
+++ b/regex-syntax/src/error.rs
@@ -1,9 +1,13 @@
-use std::cmp;
-use std::fmt;
-use std::result;
+use core::{cmp, fmt, result};
 
-use crate::ast;
-use crate::hir;
+use alloc::{
+    format,
+    string::{String, ToString},
+    vec,
+    vec::Vec,
+};
+
+use crate::{ast, hir};
 
 /// A type alias for dealing with errors returned by this crate.
 pub type Result<T> = result::Result<T, Error>;
@@ -35,6 +39,7 @@ impl From<hir::Error> for Error {
     }
 }
 
+#[cfg(feature = "std")]
 impl std::error::Error for Error {}
 
 impl fmt::Display for Error {
@@ -266,11 +271,13 @@ impl<'p> Spans<'p> {
 }
 
 fn repeat_char(c: char, count: usize) -> String {
-    ::std::iter::repeat(c).take(count).collect()
+    core::iter::repeat(c).take(count).collect()
 }
 
 #[cfg(test)]
 mod tests {
+    use alloc::string::ToString;
+
     use crate::ast::parse::Parser;
 
     fn assert_panic_message(pattern: &str, expected_msg: &str) {
diff --git a/regex-syntax/src/hir/interval.rs b/regex-syntax/src/hir/interval.rs
index d6e83f7b2..fbe772ea4 100644
--- a/regex-syntax/src/hir/interval.rs
+++ b/regex-syntax/src/hir/interval.rs
@@ -1,8 +1,6 @@
-use std::char;
-use std::cmp;
-use std::fmt::Debug;
-use std::slice;
-use std::u8;
+use core::{char, cmp, fmt::Debug, slice};
+
+use alloc::vec::Vec;
 
 use crate::unicode;
 
diff --git a/regex-syntax/src/hir/literal/mod.rs b/regex-syntax/src/hir/literal/mod.rs
index 58b8871ed..d49cffd92 100644
--- a/regex-syntax/src/hir/literal/mod.rs
+++ b/regex-syntax/src/hir/literal/mod.rs
@@ -2,11 +2,15 @@
 Provides routines for extracting literal prefixes and suffixes from an `Hir`.
 */
 
-use std::cmp;
-use std::fmt;
-use std::iter;
-use std::mem;
-use std::ops;
+use core::{cmp, iter, mem, ops};
+
+use alloc::{
+    boxed::Box,
+    format,
+    string::{String, ToString},
+    vec,
+    vec::Vec,
+};
 
 use crate::hir::{self, Hir, HirKind};
 
@@ -408,7 +412,7 @@ impl Literals {
         }
         if self.lits.is_empty() {
             let i = cmp::min(self.limit_size, bytes.len());
-            self.lits.push(Literal::new(bytes[..i].to_owned()));
+            self.lits.push(Literal::new(bytes[..i].to_vec()));
             self.lits[0].cut = i < bytes.len();
             return !self.lits[0].is_cut();
         }
@@ -465,8 +469,6 @@ impl Literals {
         cls: &hir::ClassUnicode,
         reverse: bool,
     ) -> bool {
-        use std::char;
-
         if self.class_exceeds_limits(cls_char_count(cls)) {
             return false;
         }
@@ -837,8 +839,8 @@ fn alternate_literals<F: FnMut(&Hir, &mut Literals)>(
     }
 }
 
-impl fmt::Debug for Literals {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Debug for Literals {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         f.debug_struct("Literals")
             .field("lits", &self.lits)
             .field("limit_size", &self.limit_size)
@@ -881,8 +883,8 @@ impl PartialOrd for Literal {
     }
 }
 
-impl fmt::Debug for Literal {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Debug for Literal {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         if self.is_cut() {
             write!(f, "Cut({})", escape_unicode(&self.v))
         } else {
@@ -923,7 +925,7 @@ fn position(needle: &[u8], mut haystack: &[u8]) -> Option<usize> {
 }
 
 fn escape_unicode(bytes: &[u8]) -> String {
-    let show = match ::std::str::from_utf8(bytes) {
+    let show = match core::str::from_utf8(bytes) {
         Ok(v) => v.to_string(),
         Err(_) => escape_bytes(bytes),
     };
@@ -955,7 +957,7 @@ fn escape_bytes(bytes: &[u8]) -> String {
 }
 
 fn escape_byte(byte: u8) -> String {
-    use std::ascii::escape_default;
+    use core::ascii::escape_default;
 
     let escaped: Vec<u8> = escape_default(byte).collect();
     String::from_utf8_lossy(&escaped).into_owned()
@@ -971,11 +973,15 @@ fn cls_byte_count(cls: &hir::ClassBytes) -> usize {
 
 #[cfg(test)]
 mod tests {
-    use std::fmt;
+    use alloc::{
+        string::{String, ToString},
+        vec,
+        vec::Vec,
+    };
+
+    use crate::{hir::Hir, ParserBuilder};
 
-    use super::{escape_bytes, Literal, Literals};
-    use crate::hir::Hir;
-    use crate::ParserBuilder;
+    use super::*;
 
     // To make test failures easier to read.
     #[derive(Debug, Eq, PartialEq)]
@@ -1013,8 +1019,8 @@ mod tests {
         }
     }
 
-    impl fmt::Debug for ULiteral {
-        fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+    impl core::fmt::Debug for ULiteral {
+        fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
             if self.is_cut() {
                 write!(f, "Cut({})", self.v)
             } else {
@@ -1037,11 +1043,11 @@ mod tests {
 
     #[allow(non_snake_case)]
     fn C(s: &'static str) -> ULiteral {
-        ULiteral { v: s.to_owned(), cut: true }
+        ULiteral { v: s.to_string(), cut: true }
     }
     #[allow(non_snake_case)]
     fn M(s: &'static str) -> ULiteral {
-        ULiteral { v: s.to_owned(), cut: false }
+        ULiteral { v: s.to_string(), cut: false }
     }
 
     fn prefixes(lits: &mut Literals, expr: &Hir) {
@@ -1626,7 +1632,7 @@ mod tests {
                 let given: Vec<Literal> = $given
                     .into_iter()
                     .map(|s: &str| Literal {
-                        v: s.to_owned().into_bytes(),
+                        v: s.to_string().into_bytes(),
                         cut: false,
                     })
                     .collect();
@@ -1661,7 +1667,7 @@ mod tests {
                 let given: Vec<Literal> = $given
                     .into_iter()
                     .map(|s: &str| Literal {
-                        v: s.to_owned().into_bytes(),
+                        v: s.to_string().into_bytes(),
                         cut: false,
                     })
                     .collect();
diff --git a/regex-syntax/src/hir/mod.rs b/regex-syntax/src/hir/mod.rs
index b16df20c8..2af769e92 100644
--- a/regex-syntax/src/hir/mod.rs
+++ b/regex-syntax/src/hir/mod.rs
@@ -1,18 +1,27 @@
 /*!
 Defines a high-level intermediate representation for regular expressions.
 */
-use std::char;
-use std::cmp;
-use std::fmt;
-use std::result;
-use std::u8;
 
-use crate::ast::Span;
-use crate::hir::interval::{Interval, IntervalSet, IntervalSetIter};
-use crate::unicode;
+use core::{char, cmp};
 
-pub use crate::hir::visitor::{visit, Visitor};
-pub use crate::unicode::CaseFoldError;
+use alloc::{
+    boxed::Box,
+    format,
+    string::{String, ToString},
+    vec,
+    vec::Vec,
+};
+
+use crate::{
+    ast::Span,
+    hir::interval::{Interval, IntervalSet, IntervalSetIter},
+    unicode,
+};
+
+pub use crate::{
+    hir::visitor::{visit, Visitor},
+    unicode::CaseFoldError,
+};
 
 mod interval;
 pub mod literal;
@@ -80,16 +89,17 @@ pub enum ErrorKind {
     UnicodeCaseUnavailable,
 }
 
+#[cfg(feature = "std")]
 impl std::error::Error for Error {}
 
-impl fmt::Display for Error {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for Error {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         crate::error::Formatter::from(self).fmt(f)
     }
 }
 
-impl fmt::Display for ErrorKind {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for ErrorKind {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         use self::ErrorKind::*;
 
         let msg = match *self {
@@ -197,8 +207,7 @@ impl Hir {
     /// Consumes ownership of this HIR expression and returns its underlying
     /// `HirKind`.
     pub fn into_kind(mut self) -> HirKind {
-        use std::mem;
-        mem::replace(&mut self.kind, HirKind::Empty)
+        core::mem::replace(&mut self.kind, HirKind::Empty)
     }
 
     /// Returns an empty HIR expression.
@@ -704,8 +713,8 @@ impl HirKind {
 ///
 /// This implementation uses constant stack space and heap space proportional
 /// to the size of the `Hir`.
-impl fmt::Display for Hir {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for Hir {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         use crate::hir::print::Printer;
         Printer::new().print(self, f)
     }
@@ -800,7 +809,7 @@ impl Class {
     /// Unicode oriented.
     pub fn try_case_fold_simple(
         &mut self,
-    ) -> result::Result<(), CaseFoldError> {
+    ) -> core::result::Result<(), CaseFoldError> {
         match *self {
             Class::Unicode(ref mut x) => x.try_case_fold_simple()?,
             Class::Bytes(ref mut x) => x.case_fold_simple(),
@@ -909,7 +918,7 @@ impl ClassUnicode {
     /// `unicode-case` feature is not enabled.
     pub fn try_case_fold_simple(
         &mut self,
-    ) -> result::Result<(), CaseFoldError> {
+    ) -> core::result::Result<(), CaseFoldError> {
         self.set.case_fold_simple()
     }
 
@@ -981,8 +990,8 @@ pub struct ClassUnicodeRange {
     end: char,
 }
 
-impl fmt::Debug for ClassUnicodeRange {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Debug for ClassUnicodeRange {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         let start = if !self.start.is_whitespace() && !self.start.is_control()
         {
             self.start.to_string()
@@ -1285,8 +1294,8 @@ impl ClassBytesRange {
     }
 }
 
-impl fmt::Debug for ClassBytesRange {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Debug for ClassBytesRange {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         let mut debug = f.debug_struct("ClassBytesRange");
         if self.start <= 0x7F {
             let ch = char::try_from(self.start).unwrap();
@@ -1459,7 +1468,7 @@ pub enum RepetitionRange {
 /// space but heap space proportional to the depth of the total `Hir`.
 impl Drop for Hir {
     fn drop(&mut self) {
-        use std::mem;
+        use core::mem;
 
         match *self.kind() {
             HirKind::Empty
diff --git a/regex-syntax/src/hir/print.rs b/regex-syntax/src/hir/print.rs
index 433f9bf11..63d014b1b 100644
--- a/regex-syntax/src/hir/print.rs
+++ b/regex-syntax/src/hir/print.rs
@@ -2,11 +2,16 @@
 This module provides a regular expression printer for `Hir`.
 */
 
-use std::fmt;
+use core::fmt;
 
-use crate::hir::visitor::{self, Visitor};
-use crate::hir::{self, Hir, HirKind};
-use crate::is_meta_character;
+use crate::{
+    hir::{
+        self,
+        visitor::{self, Visitor},
+        Hir, HirKind,
+    },
+    is_meta_character,
+};
 
 /// A builder for constructing a printer.
 ///
@@ -235,9 +240,12 @@ impl<W: fmt::Write> Writer<W> {
 
 #[cfg(test)]
 mod tests {
-    use super::Printer;
+    use alloc::string::String;
+
     use crate::ParserBuilder;
 
+    use super::*;
+
     fn roundtrip(given: &str, expected: &str) {
         roundtrip_with(|b| b, given, expected);
     }
diff --git a/regex-syntax/src/hir/translate.rs b/regex-syntax/src/hir/translate.rs
index 988384ede..b4338bc94 100644
--- a/regex-syntax/src/hir/translate.rs
+++ b/regex-syntax/src/hir/translate.rs
@@ -2,14 +2,17 @@
 Defines a translator that converts an `Ast` to an `Hir`.
 */
 
-use std::cell::{Cell, RefCell};
-use std::result;
+use core::cell::{Cell, RefCell};
 
-use crate::ast::{self, Ast, Span, Visitor};
-use crate::hir::{self, Error, ErrorKind, Hir};
-use crate::unicode::{self, ClassQuery};
+use alloc::{boxed::Box, string::ToString, vec, vec::Vec};
 
-type Result<T> = result::Result<T, Error>;
+use crate::{
+    ast::{self, Ast, Span, Visitor},
+    hir::{self, Error, ErrorKind, Hir},
+    unicode::{self, ClassQuery},
+};
+
+type Result<T> = core::result::Result<T, Error>;
 
 /// A builder for constructing an AST->HIR translator.
 #[derive(Clone, Debug)]
@@ -1119,12 +1122,13 @@ fn ascii_class_as_chars(
 
 #[cfg(test)]
 mod tests {
-    use crate::ast::parse::ParserBuilder;
-    use crate::ast::{self, Ast, Position, Span};
-    use crate::hir::{self, Hir, HirKind};
-    use crate::unicode::{self, ClassQuery};
+    use crate::{
+        ast::{self, parse::ParserBuilder, Ast, Position, Span},
+        hir::{self, Hir, HirKind},
+        unicode::{self, ClassQuery},
+    };
 
-    use super::{ascii_class, ascii_class_as_chars, TranslatorBuilder};
+    use super::*;
 
     // We create these errors to compare with real hir::Errors in the tests.
     // We define equality between TestError and hir::Error to disregard the
diff --git a/regex-syntax/src/hir/visitor.rs b/regex-syntax/src/hir/visitor.rs
index 4f5a70909..97771d92f 100644
--- a/regex-syntax/src/hir/visitor.rs
+++ b/regex-syntax/src/hir/visitor.rs
@@ -1,3 +1,5 @@
+use alloc::{vec, vec::Vec};
+
 use crate::hir::{self, Hir, HirKind};
 
 /// A trait for visiting the high-level IR (HIR) in depth first order.
diff --git a/regex-syntax/src/lib.rs b/regex-syntax/src/lib.rs
index 1dfb38af3..287b3417c 100644
--- a/regex-syntax/src/lib.rs
+++ b/regex-syntax/src/lib.rs
@@ -116,6 +116,11 @@ match semantics of a regular expression.
 
 The following features are available:
 
+* **std** -
+  Enables support for the standard library. This feature is enabled by default.
+  When disabled, only `core` and `alloc` are used. Otherwise, enabling `std`
+  generally just enables `std::error::Error` trait impls for the various error
+  types.
 * **unicode** -
   Enables all Unicode features. This feature is enabled by default, and will
   always cover all Unicode features, even if more are added in the future.
@@ -154,13 +159,23 @@ The following features are available:
   `\p{sb=ATerm}`.
 */
 
+#![forbid(unsafe_code)]
 #![deny(missing_docs)]
 #![warn(missing_debug_implementations)]
-#![forbid(unsafe_code)]
+#![no_std]
+
+#[cfg(any(test, feature = "std"))]
+extern crate std;
 
-pub use crate::error::{Error, Result};
-pub use crate::parser::{Parser, ParserBuilder};
-pub use crate::unicode::UnicodeWordError;
+extern crate alloc;
+
+pub use crate::{
+    error::{Error, Result},
+    parser::{Parser, ParserBuilder},
+    unicode::UnicodeWordError,
+};
+
+use alloc::string::String;
 
 pub mod ast;
 mod either;
@@ -248,7 +263,7 @@ pub fn is_word_character(c: char) -> bool {
 /// returns an error.
 pub fn try_is_word_character(
     c: char,
-) -> std::result::Result<bool, UnicodeWordError> {
+) -> core::result::Result<bool, UnicodeWordError> {
     unicode::is_word_character(c)
 }
 
@@ -265,6 +280,8 @@ pub fn is_word_byte(c: u8) -> bool {
 
 #[cfg(test)]
 mod tests {
+    use alloc::string::ToString;
+
     use super::*;
 
     #[test]
diff --git a/regex-syntax/src/parser.rs b/regex-syntax/src/parser.rs
index ded95b280..93df72279 100644
--- a/regex-syntax/src/parser.rs
+++ b/regex-syntax/src/parser.rs
@@ -1,7 +1,4 @@
-use crate::ast;
-use crate::hir;
-
-use crate::Result;
+use crate::{ast, hir, Result};
 
 /// A builder for a regular expression parser.
 ///
diff --git a/regex-syntax/src/unicode.rs b/regex-syntax/src/unicode.rs
index 0b716f5e6..1689681fa 100644
--- a/regex-syntax/src/unicode.rs
+++ b/regex-syntax/src/unicode.rs
@@ -1,11 +1,12 @@
-use std::error;
-use std::fmt;
-use std::result;
+use alloc::{
+    string::{String, ToString},
+    vec::Vec,
+};
 
 use crate::hir;
 
 /// A type alias for errors specific to Unicode handling of classes.
-pub type Result<T> = result::Result<T, Error>;
+pub type Result<T> = core::result::Result<T, Error>;
 
 /// An inclusive range of codepoints from a generated file (hence the static
 /// lifetime).
@@ -25,7 +26,7 @@ pub enum Error {
 }
 
 /// A type alias for errors specific to Unicode case folding.
-pub type FoldResult<T> = result::Result<T, CaseFoldError>;
+pub type FoldResult<T> = core::result::Result<T, CaseFoldError>;
 
 /// An error that occurs when Unicode-aware simple case folding fails.
 ///
@@ -35,10 +36,11 @@ pub type FoldResult<T> = result::Result<T, CaseFoldError>;
 #[derive(Debug)]
 pub struct CaseFoldError(());
 
-impl error::Error for CaseFoldError {}
+#[cfg(feature = "std")]
+impl std::error::Error for CaseFoldError {}
 
-impl fmt::Display for CaseFoldError {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for CaseFoldError {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         write!(
             f,
             "Unicode-aware case folding is not available \
@@ -55,10 +57,11 @@ impl fmt::Display for CaseFoldError {
 #[derive(Debug)]
 pub struct UnicodeWordError(());
 
-impl error::Error for UnicodeWordError {}
+#[cfg(feature = "std")]
+impl std::error::Error for UnicodeWordError {}
 
-impl fmt::Display for UnicodeWordError {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+impl core::fmt::Display for UnicodeWordError {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         write!(
             f,
             "Unicode-aware \\w class is not available \
@@ -80,21 +83,24 @@ impl fmt::Display for UnicodeWordError {
 /// This returns an error if the Unicode case folding tables are not available.
 pub fn simple_fold(
     c: char,
-) -> FoldResult<result::Result<impl Iterator<Item = char>, Option<char>>> {
+) -> FoldResult<core::result::Result<impl Iterator<Item = char>, Option<char>>>
+{
     #[cfg(not(feature = "unicode-case"))]
     fn imp(
         _: char,
-    ) -> FoldResult<result::Result<impl Iterator<Item = char>, Option<char>>>
-    {
-        use std::option::IntoIter;
-        Err::<result::Result<IntoIter<char>, _>, _>(CaseFoldError(()))
+    ) -> FoldResult<
+        core::result::Result<impl Iterator<Item = char>, Option<char>>,
+    > {
+        use core::option::IntoIter;
+        Err::<core::result::Result<IntoIter<char>, _>, _>(CaseFoldError(()))
     }
 
     #[cfg(feature = "unicode-case")]
     fn imp(
         c: char,
-    ) -> FoldResult<result::Result<impl Iterator<Item = char>, Option<char>>>
-    {
+    ) -> FoldResult<
+        core::result::Result<impl Iterator<Item = char>, Option<char>>,
+    > {
         use crate::unicode_tables::case_folding_simple::CASE_FOLDING_SIMPLE;
 
         Ok(CASE_FOLDING_SIMPLE
@@ -130,8 +136,9 @@ pub fn contains_simple_case_mapping(
 
     #[cfg(feature = "unicode-case")]
     fn imp(start: char, end: char) -> FoldResult<bool> {
+        use core::cmp::Ordering;
+
         use crate::unicode_tables::case_folding_simple::CASE_FOLDING_SIMPLE;
-        use std::cmp::Ordering;
 
         assert!(start <= end);
         Ok(CASE_FOLDING_SIMPLE
@@ -407,17 +414,17 @@ pub fn hir_class(ranges: &[(char, char)]) -> hir::ClassUnicode {
 /// Returns true only if the given codepoint is in the `\w` character class.
 ///
 /// If the `unicode-perl` feature is not enabled, then this returns an error.
-pub fn is_word_character(c: char) -> result::Result<bool, UnicodeWordError> {
+pub fn is_word_character(
+    c: char,
+) -> core::result::Result<bool, UnicodeWordError> {
     #[cfg(not(feature = "unicode-perl"))]
-    fn imp(_: char) -> result::Result<bool, UnicodeWordError> {
+    fn imp(_: char) -> core::result::Result<bool, UnicodeWordError> {
         Err(UnicodeWordError(()))
     }
 
     #[cfg(feature = "unicode-perl")]
-    fn imp(c: char) -> result::Result<bool, UnicodeWordError> {
-        use crate::is_word_byte;
-        use crate::unicode_tables::perl_word::PERL_WORD;
-        use std::cmp::Ordering;
+    fn imp(c: char) -> core::result::Result<bool, UnicodeWordError> {
+        use crate::{is_word_byte, unicode_tables::perl_word::PERL_WORD};
 
         // MSRV(1.59): Use 'u8::try_from(c)' instead.
         if u8::try_from(u32::from(c)).map_or(false, is_word_byte) {
@@ -425,6 +432,8 @@ pub fn is_word_character(c: char) -> result::Result<bool, UnicodeWordError> {
         }
         Ok(PERL_WORD
             .binary_search_by(|&(start, end)| {
+                use core::cmp::Ordering;
+
                 if start <= c && c <= end {
                     Ordering::Equal
                 } else if start > c {
@@ -583,7 +592,7 @@ fn property_set(
 fn ages(canonical_age: &str) -> Result<impl Iterator<Item = Range>> {
     #[cfg(not(feature = "unicode-age"))]
     fn imp(_: &str) -> Result<impl Iterator<Item = Range>> {
-        use std::option::IntoIter;
+        use core::option::IntoIter;
         Err::<IntoIter<Range>, _>(Error::PropertyNotFound)
     }
 
@@ -884,10 +893,7 @@ fn symbolic_name_normalize_bytes(slice: &mut [u8]) -> &mut [u8] {
 
 #[cfg(test)]
 mod tests {
-    use super::{
-        contains_simple_case_mapping, simple_fold, symbolic_name_normalize,
-        symbolic_name_normalize_bytes,
-    };
+    use super::*;
 
     #[cfg(feature = "unicode-case")]
     fn simple_fold_ok(c: char) -> impl Iterator<Item = char> {
@@ -911,23 +917,23 @@ mod tests {
     #[cfg(feature = "unicode-case")]
     fn simple_fold_k() {
         let xs: Vec<char> = simple_fold_ok('k').collect();
-        assert_eq!(xs, vec!['K', 'K']);
+        assert_eq!(xs, alloc::vec!['K', 'K']);
 
         let xs: Vec<char> = simple_fold_ok('K').collect();
-        assert_eq!(xs, vec!['k', 'K']);
+        assert_eq!(xs, alloc::vec!['k', 'K']);
 
         let xs: Vec<char> = simple_fold_ok('K').collect();
-        assert_eq!(xs, vec!['K', 'k']);
+        assert_eq!(xs, alloc::vec!['K', 'k']);
     }
 
     #[test]
     #[cfg(feature = "unicode-case")]
     fn simple_fold_a() {
         let xs: Vec<char> = simple_fold_ok('a').collect();
-        assert_eq!(xs, vec!['A']);
+        assert_eq!(xs, alloc::vec!['A']);
 
         let xs: Vec<char> = simple_fold_ok('A').collect();
-        assert_eq!(xs, vec!['a']);
+        assert_eq!(xs, alloc::vec!['a']);
     }
 
     #[test]
diff --git a/regex-syntax/src/utf8.rs b/regex-syntax/src/utf8.rs
index b00cd7dba..a75a8afa8 100644
--- a/regex-syntax/src/utf8.rs
+++ b/regex-syntax/src/utf8.rs
@@ -80,12 +80,9 @@ I also got the idea from
 which uses it for executing automata on their term index.
 */
 
-#![deny(missing_docs)]
+use core::{char, fmt, iter::FusedIterator, slice};
 
-use std::char;
-use std::fmt;
-use std::iter::FusedIterator;
-use std::slice;
+use alloc::{vec, vec::Vec};
 
 const MAX_UTF8_BYTES: usize = 4;
 
@@ -457,7 +454,9 @@ fn max_scalar_value(nbytes: usize) -> u32 {
 
 #[cfg(test)]
 mod tests {
-    use std::char;
+    use core::char;
+
+    use alloc::{vec, vec::Vec};
 
     use crate::utf8::{Utf8Range, Utf8Sequences};
 
diff --git a/regex-syntax/test b/regex-syntax/test
index 4b1b9fb1a..d03db94b4 100755
--- a/regex-syntax/test
+++ b/regex-syntax/test
@@ -7,6 +7,7 @@ echo "===== DEFAULT FEATURES ==="
 cargo test
 
 features=(
+    std
     unicode
     unicode-age
     unicode-bool