diff --git a/README.md b/README.md
index 15e2c7df..c530e688 100644
--- a/README.md
+++ b/README.md
@@ -1410,6 +1410,26 @@ The example below solves the problem of indenting function application in OCaml
 )
 ```
 
+### `@lower_case`/`@upper_case`
+(Don't) Capitalize all of the text in the matched node.
+Use this with care in languages that are case sensitive.
+
+```scheme
+; example for SQL, since that's where this makes sense.
+; I am using the grammar linked below
+; https://github.com/DerekStride/tree-sitter-sql/tree/main
+
+; make keywords select,from lowercase.
+(
+ [
+  (keyword_select)
+  (keyword_from)
+ ] @lower_case
+)
+; make keyword WHERE uppercase
+(keyword_where) @upper_case
+```
+
 ### `#query_name!`
 
 When the logging verbosity is set to `-vv` or higher, Topiary outputs information about which queries are matched, for instance:
diff --git a/topiary-cli/Cargo.toml b/topiary-cli/Cargo.toml
index ad0bb884..ad4c8515 100644
--- a/topiary-cli/Cargo.toml
+++ b/topiary-cli/Cargo.toml
@@ -76,7 +76,7 @@ contributed = [
 
 # Excluded by default
 experimental = [
-  "rust",
+  "rust"
 ]
 
 bash = ["topiary-config/bash", "topiary-queries/bash"]
diff --git a/topiary-config/Cargo.toml b/topiary-config/Cargo.toml
index 94f92725..f2ecc0eb 100644
--- a/topiary-config/Cargo.toml
+++ b/topiary-config/Cargo.toml
@@ -49,6 +49,7 @@ ocamllex = []
 rust = []
 toml = []
 tree_sitter_query = []
+sql = []
 
 # This a convenience for the sake of downstream applications which don't
 # wish to cherry-pick grammars (e.g., the playground)
@@ -62,5 +63,5 @@ all = [
   "ocamllex",
   "rust",
   "toml",
-  "tree_sitter_query"
+  "tree_sitter_query",
 ]
diff --git a/topiary-core/src/atom_collection.rs b/topiary-core/src/atom_collection.rs
index a1e6d64f..50af01be 100644
--- a/topiary-core/src/atom_collection.rs
+++ b/topiary-core/src/atom_collection.rs
@@ -1,5 +1,6 @@
 use std::{
     borrow::Cow,
+    cmp::max,
     collections::{HashMap, HashSet},
     mem,
     ops::Deref,
@@ -8,7 +9,8 @@ use std::{
 use topiary_tree_sitter_facade::Node;
 
 use crate::{
-    tree_sitter::NodeExt, Atom, FormatterError, FormatterResult, ScopeCondition, ScopeInformation,
+    tree_sitter::NodeExt, Atom, FormatterError, FormatterResult, HowCapitalize, ScopeCondition,
+    ScopeInformation,
 };
 
 /// A struct that holds sets of node IDs that have line breaks before or after them.
@@ -278,6 +280,14 @@ impl AtomCollection {
                 self.prepend(Atom::DeleteBegin, node, predicates);
                 self.append(Atom::DeleteEnd, node, predicates);
             }
+            "upper_case" => {
+                self.prepend(Atom::CaseBegin(HowCapitalize::UpperCase), node, predicates);
+                self.append(Atom::CaseEnd, node, predicates);
+            }
+            "lower_case" => {
+                self.prepend(Atom::CaseBegin(HowCapitalize::LowerCase), node, predicates);
+                self.append(Atom::CaseEnd, node, predicates);
+            }
             // Scope manipulation
             "prepend_begin_scope" => {
                 self.prepend(
@@ -538,6 +548,7 @@ impl AtomCollection {
                 original_position: node.start_position().into(),
                 single_line_no_indent: false,
                 multi_line_indent_all: false,
+                how_capitalize: HowCapitalize::Pass,
             });
             // Mark all sub-nodes as having this node as a "leaf parent"
             self.mark_leaf_parent(node, node.id());
@@ -591,6 +602,7 @@ impl AtomCollection {
         self.append.entry(target_node.id()).or_default().push(atom);
     }
 
+    // fn capitalize(&mut self,
     /// Expands a softline atom to a hardline, space or empty atom depending on
     /// if we are in a multiline context or not.
     ///
@@ -883,6 +895,63 @@ impl AtomCollection {
         }
     }
 
+    fn post_process_capitalization(&mut self) {
+        let mut case_context: Vec<HowCapitalize> = Vec::new();
+        for atom in &mut self.atoms {
+            match atom {
+                Atom::CaseBegin(case) => {
+                    match case {
+                        HowCapitalize::UpperCase => {
+                            case_context.push(HowCapitalize::UpperCase);
+                            *atom = Atom::Empty;
+                            /*
+                            if lower_case_level > 0 {
+                                panic!("Cannot use @lower_case inside of a node captured by @upper_case!");
+                            }*/
+                        }
+                        HowCapitalize::LowerCase => {
+                            case_context.push(HowCapitalize::LowerCase);
+                            *atom = Atom::Empty;
+                            /*
+                            if upper_case_level > 0 {
+                                panic!("Cannot use @upper_case inside of a node captured by @lower_case!");
+                            }*/
+                        }
+                        _ => {}
+                    }
+                }
+                Atom::CaseEnd => {
+                    case_context.pop();
+                    *atom = Atom::Empty;
+                }
+                _ => match atom {
+                    Atom::Leaf {
+                        content,
+                        id,
+                        original_position,
+                        single_line_no_indent,
+                        multi_line_indent_all,
+                        ..
+                    } => {
+                        // TODO don't be stupid with derefs
+                        *atom = Atom::Leaf {
+                            content: (*content).to_string(),
+                            id: *id,
+                            original_position: *original_position,
+                            single_line_no_indent: *single_line_no_indent,
+                            multi_line_indent_all: *multi_line_indent_all,
+                            how_capitalize: case_context
+                                .last()
+                                .unwrap_or(&HowCapitalize::Pass)
+                                .clone(),
+                        }
+                    }
+                    _ => {}
+                },
+            }
+        }
+    }
+
     /// This function merges the spaces, new lines and blank lines.
     /// If there are several tokens of different kind one after the other,
     /// the blank line is kept over the new line which itself is kept over the space.
@@ -890,6 +959,7 @@ impl AtomCollection {
     pub fn post_process(&mut self) {
         self.post_process_scopes();
         self.post_process_deletes();
+        self.post_process_capitalization();
         self.post_process_inner();
 
         // We have taken care of spaces following an antispace. Now fix the
diff --git a/topiary-core/src/lib.rs b/topiary-core/src/lib.rs
index d03a38b9..2918ed37 100644
--- a/topiary-core/src/lib.rs
+++ b/topiary-core/src/lib.rs
@@ -38,6 +38,13 @@ pub struct ScopeInformation {
     scope_id: String,
 }
 
+#[derive(Clone, Debug, Default, Eq, PartialEq)]
+pub enum HowCapitalize {
+    UpperCase,
+    LowerCase,
+    #[default]
+    Pass,
+}
 /// An atom represents a small piece of the output. We turn Tree-sitter nodes
 /// into atoms, and we add white-space atoms where appropriate. The final list
 /// of atoms is rendered to the output.
@@ -68,6 +75,7 @@ pub enum Atom {
         single_line_no_indent: bool,
         // if the leaf is multi-line, each line will be indented, not just the first
         multi_line_indent_all: bool,
+        how_capitalize: HowCapitalize,
     },
     /// Represents a literal string, such as a semicolon.
     Literal(String),
@@ -86,6 +94,9 @@ pub enum Atom {
     // it might happen that it contains several leaves.
     DeleteBegin,
     DeleteEnd,
+
+    CaseBegin(HowCapitalize),
+    CaseEnd,
     /// Indicates the beginning of a scope, use in combination with the
     /// ScopedSoftlines and ScopedConditionals below.
     ScopeBegin(ScopeInformation),
diff --git a/topiary-core/src/pretty.rs b/topiary-core/src/pretty.rs
index 8c94c285..d09379f8 100644
--- a/topiary-core/src/pretty.rs
+++ b/topiary-core/src/pretty.rs
@@ -4,7 +4,7 @@
 
 use std::fmt::Write;
 
-use crate::{Atom, FormatterError, FormatterResult};
+use crate::{Atom, FormatterError, FormatterResult, HowCapitalize};
 
 /// Renders a slice of Atoms into an owned string.
 /// The indent &str is used when an `Atom::IdentStart` is encountered.
@@ -44,6 +44,7 @@ pub fn render(atoms: &[Atom], indent: &str) -> FormatterResult<String> {
                 original_position,
                 single_line_no_indent,
                 multi_line_indent_all,
+                how_capitalize,
                 ..
             } => {
                 if *single_line_no_indent {
@@ -51,10 +52,9 @@ pub fn render(atoms: &[Atom], indent: &str) -> FormatterResult<String> {
                     // as a `Hardline` in the atom stream.
                     writeln!(buffer)?;
                 }
-
                 let content = content.trim_end_matches('\n');
 
-                let content = if *multi_line_indent_all {
+                let mut content = if *multi_line_indent_all {
                     let cursor = current_column(&buffer) as i32;
 
                     // original_position is 1-based
@@ -71,7 +71,15 @@ pub fn render(atoms: &[Atom], indent: &str) -> FormatterResult<String> {
                 } else {
                     content.into()
                 };
-
+                match how_capitalize {
+                    HowCapitalize::UpperCase => {
+                        content = content.to_uppercase();
+                    }
+                    HowCapitalize::LowerCase => {
+                        content = content.to_lowercase();
+                    }
+                    _ => {}
+                }
                 write!(buffer, "{}", content)?;
             }