diff --git a/src/openhuman/memory_tree/score/extract/types.rs b/src/openhuman/memory_tree/score/extract/types.rs
index 5458238ec..1ad65b35b 100644
--- a/src/openhuman/memory_tree/score/extract/types.rs
+++ b/src/openhuman/memory_tree/score/extract/types.rs
@@ -153,11 +153,13 @@ impl ExtractedEntities {
 
     /// Count of unique `(kind, text)` pairs, case-insensitive. Used as a scoring signal.
     pub fn unique_entity_count(&self) -> usize {
-        use std::collections::BTreeSet;
+        use std::collections::HashSet;
+        // Only the cardinality matters, so a hash set counts the distinct
+        // `(kind, text)` pairs without paying for ordered insertion.
         self.entities
             .iter()
             .map(|e| (e.kind, e.text.to_lowercase()))
-            .collect::<BTreeSet<_>>()
+            .collect::<HashSet<_>>()
             .len()
     }
 
@@ -172,8 +174,11 @@ impl ExtractedEntities {
     /// The reason from whichever side won the max wins; if they tied or
     /// both are absent, the non-empty one (if any) is kept.
     pub fn merge(&mut self, other: ExtractedEntities) {
-        use std::collections::BTreeSet;
-        let mut seen: BTreeSet<(EntityKind, String, u32)> = self
+        use std::collections::HashSet;
+        // Both sets are membership-only dedup guards — the surviving order is
+        // the existing `Vec` push order, never the set's — so a hash set keeps
+        // the merge result identical while dropping the ordered-key overhead.
+        let mut seen: HashSet<(EntityKind, String, u32)> = self
             .entities
             .iter()
             .map(|e| (e.kind, e.text.to_lowercase(), e.span_start))
@@ -184,8 +189,7 @@ impl ExtractedEntities {
                 self.entities.push(e);
             }
         }
-        let mut topic_seen: BTreeSet<String> =
-            self.topics.iter().map(|t| t.label.clone()).collect();
+        let mut topic_seen: HashSet<String> = self.topics.iter().map(|t| t.label.clone()).collect();
         for t in other.topics {
             if topic_seen.insert(t.label.clone()) {
                 self.topics.push(t);
diff --git a/src/openhuman/memory_tree/score/signals/unique_words.rs b/src/openhuman/memory_tree/score/signals/unique_words.rs
index 0ae1b6e3d..8436d62f2 100644
--- a/src/openhuman/memory_tree/score/signals/unique_words.rs
+++ b/src/openhuman/memory_tree/score/signals/unique_words.rs
@@ -20,7 +20,10 @@ pub const MIN_TOTAL_WORDS: usize = 5;
 /// - Linear in between
 pub fn score(text: &str) -> f32 {
     let mut total: usize = 0;
-    let mut uniq: std::collections::BTreeSet<String> = std::collections::BTreeSet::new();
+    // Only `uniq.len()` is ever read — never the iteration order — so a hash set
+    // gives the identical type-token ratio with O(1) inserts instead of the
+    // ordered set's O(log n) String comparisons per word.
+    let mut uniq: std::collections::HashSet<String> = std::collections::HashSet::new();
 
     for raw in text.split_whitespace() {
         let w: String = raw