diff --git a/.changeset/range-bound-escapes.md b/.changeset/range-bound-escapes.md
new file mode 100644
index 0000000000..39b9e79cec
--- /dev/null
+++ b/.changeset/range-bound-escapes.md
@@ -0,0 +1,12 @@
+---
+'@hyperdx/common-utils': patch
+'@hyperdx/api': patch
+'@hyperdx/app': patch
+---
+
+fix: decode escaped characters in Lucene range bounds
+
+A range bound with an escaped colon, such as
+`Timestamp:[2024-01-01T10\:00\:00 TO *]`, sent the internal placeholder to
+ClickHouse (`'2024-01-01T10HDX_COLON00HDX_COLON00'`) and the query failed. Range
+bounds are now decoded the same way as other search terms.
diff --git a/packages/common-utils/src/__tests__/queryParser.test.ts b/packages/common-utils/src/__tests__/queryParser.test.ts
index c3f897b3db..b6bb091786 100644
--- a/packages/common-utils/src/__tests__/queryParser.test.ts
+++ b/packages/common-utils/src/__tests__/queryParser.test.ts
@@ -842,6 +842,14 @@ describe('CustomSchemaSQLSerializerV2 - range bounds', () => {
       lucene: 'Timestamp:{2024-01-01 TO 2024-06-01}',
       sql: "((Timestamp > '2024-01-01' AND Timestamp < '2024-06-01'))",
     },
+    {
+      lucene: 'Timestamp:[2024-01-01T10\\:00\\:00 TO *]',
+      sql: "((Timestamp >= '2024-01-01T10:00:00'))",
+    },
+    {
+      lucene: 'Host:[localhost:3000 TO localhost:3999]',
+      sql: "((Host BETWEEN 'localhost:3000' AND 'localhost:3999'))",
+    },
     {
       lucene: 'LogAttributes.duration_ms:{100 TO 500}',
       sql: "((`LogAttributes`['duration_ms'] > 100 AND `LogAttributes`['duration_ms'] < 500 AND indexHint(mapContains(`LogAttributes`, 'duration_ms'))))",
diff --git a/packages/common-utils/src/queryParser.ts b/packages/common-utils/src/queryParser.ts
index 61326a599a..37cc098afa 100644
--- a/packages/common-utils/src/queryParser.ts
+++ b/packages/common-utils/src/queryParser.ts
@@ -2083,8 +2083,8 @@ async function nodeTerm(
     const rangedTerm = node;
     return serializer.range(
       field,
-      rangedTerm.term_min,
-      rangedTerm.term_max,
+      decodeSpecialTokens(rangedTerm.term_min),
+      decodeSpecialTokens(rangedTerm.term_max),
       isNegatedField,
       context,
       rangedTerm.inclusive,