diff --git a/pkg/database/query-for-ids.go b/pkg/database/query-for-ids.go
index 5ed9017..0244ee3 100644
--- a/pkg/database/query-for-ids.go
+++ b/pkg/database/query-for-ids.go
@@ -13,7 +13,9 @@ import (
 
 // QueryForIds retrieves a list of IdPkTs based on the provided filter.
 // It supports filtering by ranges and tags but disallows filtering by Ids.
-// Results are sorted by timestamp in reverse chronological order.
+// Results are sorted by timestamp in reverse chronological order by default.
+// When a search query is present, results are ranked by a 50/50 blend of
+// match count (how many distinct search terms matched) and recency.
 // Returns an error if the filter contains Ids or if any operation fails.
 func (d *D) QueryForIds(c context.Context, f *filter.F) (
 	idPkTs []*store.IdPkTs, err error,
@@ -29,6 +31,9 @@ func (d *D) QueryForIds(c context.Context, f *filter.F) (
 	}
 	var results []*store.IdPkTs
 	var founds []*types.Uint40
+	// When searching, we want to count how many index ranges (search terms)
+	// matched each note. We'll track counts by serial.
+	counts := make(map[uint64]int)
 	for _, idx := range idxs {
 		if founds, err = d.GetSerialsByRange(idx); chk.E(err) {
 			return
@@ -37,6 +42,12 @@ func (d *D) QueryForIds(c context.Context, f *filter.F) (
 		if tmp, err = d.GetFullIdPubkeyBySerials(founds); chk.E(err) {
 			return
 		}
+		// If this query is driven by Search terms, increment count per serial
+		if len(f.Search) > 0 {
+			for _, v := range tmp {
+				counts[v.Ser]++
+			}
+		}
 		results = append(results, tmp...)
 	}
 	// deduplicate in case this somehow happened (such as two or more
@@ -48,12 +59,58 @@ func (d *D) QueryForIds(c context.Context, f *filter.F) (
 			idPkTs = append(idPkTs, idpk)
 		}
 	}
-	// sort results by timestamp in reverse chronological order
-	sort.Slice(
-		idPkTs, func(i, j int) bool {
-			return idPkTs[i].Ts > idPkTs[j].Ts
-		},
-	)
+
+	if len(f.Search) == 0 {
+		// No search query: sort by timestamp in reverse chronological order
+		sort.Slice(
+			idPkTs, func(i, j int) bool {
+				return idPkTs[i].Ts > idPkTs[j].Ts
+			},
+		)
+	} else {
+		// Search query present: blend match count relevance with recency (50/50)
+		// Normalize both match count and timestamp to [0,1] and compute score.
+		var maxCount int
+		var minTs, maxTs int64
+		if len(idPkTs) > 0 {
+			minTs, maxTs = idPkTs[0].Ts, idPkTs[0].Ts
+		}
+		for _, v := range idPkTs {
+			if c := counts[v.Ser]; c > maxCount {
+				maxCount = c
+			}
+			if v.Ts < minTs {
+				minTs = v.Ts
+			}
+			if v.Ts > maxTs {
+				maxTs = v.Ts
+			}
+		}
+		// Precompute denominator to avoid div-by-zero
+		tsSpan := maxTs - minTs
+		if tsSpan <= 0 {
+			tsSpan = 1
+		}
+		if maxCount <= 0 {
+			maxCount = 1
+		}
+		sort.Slice(
+			idPkTs, func(i, j int) bool {
+				ci := float64(counts[idPkTs[i].Ser]) / float64(maxCount)
+				cj := float64(counts[idPkTs[j].Ser]) / float64(maxCount)
+				ai := float64(idPkTs[i].Ts-minTs) / float64(tsSpan)
+				aj := float64(idPkTs[j].Ts-minTs) / float64(tsSpan)
+				si := 0.5*ci + 0.5*ai
+				sj := 0.5*cj + 0.5*aj
+				if si == sj {
+					// tie-break by recency
+					return idPkTs[i].Ts > idPkTs[j].Ts
+				}
+				return si > sj
+			},
+		)
+	}
+
 	if f.Limit != nil && len(idPkTs) > int(*f.Limit) {
 		idPkTs = idPkTs[:*f.Limit]
 	}
diff --git a/pkg/version/version b/pkg/version/version
index 7965b36..7f7306b 100644
--- a/pkg/version/version
+++ b/pkg/version/version
@@ -1 +1 @@
-v0.9.0
\ No newline at end of file
+v0.9.1
\ No newline at end of file