diff --git a/src/letterset.rs b/src/letterset.rs
index 3b51fe5..bf9838c 100644
--- a/src/letterset.rs
+++ b/src/letterset.rs
@@ -1,6 +1,6 @@
 use std::{fmt, ops::Deref};
 
-#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Default, Debug)]
+#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Default, Debug, Hash)]
 pub struct LetterSet(u32);
 
 /// Immutable set of English letters
@@ -60,6 +60,10 @@ impl LetterSet {
         self.0.count_ones() as u8
     }
 
+    pub const fn is_empty(self) -> bool {
+        self.0 == 0
+    }
+
     const fn letter_index(letter: u8) -> Option<u8> {
         if letter.is_ascii_alphabetic() {
             let lower = letter.to_ascii_lowercase();
@@ -98,6 +102,15 @@ where
     }
 }
 
+impl<T> From<T> for LetterSet
+where
+    T: AsRef<[u8]>,
+{
+    fn from(value: T) -> Self {
+        Self::from_slice(value.as_ref())
+    }
+}
+
 #[cfg(test)]
 mod test {
     use super::*;
@@ -181,4 +194,13 @@ mod test {
         assert_eq!(A.add(b'b'), AB);
         assert_eq!(AB.add(b'c'), ABC);
     }
+
+    #[test]
+    fn is_empty() {
+        assert!(LetterSet::EMPTY.is_empty());
+        assert!(!LetterSet::FULL.is_empty());
+        assert!(!A.is_empty());
+        assert!(!AB.is_empty());
+        assert!(!BC.is_empty());
+    }
 }
diff --git a/src/main.rs b/src/main.rs
index b27f6f6..04d57c2 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -1,5 +1,50 @@
+use std::{
+    collections::{HashMap, hash_map::Entry},
+    io::{self, BufRead, BufReader},
+};
+
+use crate::letterset::LetterSet;
+
 mod letterset;
 
 fn main() {
-    println!("Hello, world!");
+    let dictionary: Vec<String> = BufReader::new(io::stdin().lock())
+        .lines()
+        .collect::<Result<_, _>>()
+        .unwrap();
+
+    let mut min_lengths: HashMap<LetterSet, u16> = HashMap::new();
+    let min_length = solve(dictionary.as_slice(), &mut min_lengths, LetterSet::FULL);
+    println!("{min_length}");
+}
+
+// length doesn't include spaces. i think that's okay
+fn solve(
+    dictionary: &[String],
+    min_lengths: &mut HashMap<LetterSet, u16>,
+    remaining: LetterSet,
+) -> u16 {
+    if remaining.is_empty() {
+        return 0;
+    }
+
+    if let Some(min) = min_lengths.get(&remaining) {
+        return *min;
+    }
+
+    let min_length = dictionary
+        .iter()
+        .filter(|word| remaining != remaining.difference(LetterSet::from(word.as_bytes())))
+        .map(|word| {
+            (u16::try_from(word.len()).unwrap()).saturating_add(solve(
+                dictionary,
+                min_lengths,
+                remaining.difference(LetterSet::from(word)),
+            ))
+        })
+        .min()
+        .unwrap_or(u16::MAX);
+
+    min_lengths.insert(remaining, min_length);
+    min_length
 }