diff --git a/tradein-mvp/backend/data/sql/097_index_hygiene.sql b/tradein-mvp/backend/data/sql/097_index_hygiene.sql
new file mode 100644
index 00000000..5924bb2f
--- /dev/null
+++ b/tradein-mvp/backend/data/sql/097_index_hygiene.sql
@@ -0,0 +1,93 @@
+-- 097_index_hygiene.sql
+-- #730 index hygiene: drop exact-duplicate + redundant indexes;
+-- de-partial is_active indexes. 87-unused sweep deferred (needs live pg_stat snapshot).
+--
+-- Context (Issue #730):
+--   Duplicate drops: 5 named indexes that are exact duplicates of a surviving twin
+--   or covered by a UNIQUE constraint (all confirmed from migration history below).
+--   De-partial: 5 indexes on listings using WHERE is_active = true; since is_active
+--   is 100% true in production the partial predicate saves nothing — a full index
+--   strictly covers the same rows with no plan regression.
+--
+-- Idempotency:
+--   All DROP INDEX use IF EXISTS; all CREATE INDEX use IF NOT EXISTS.
+--   Safe to re-run.
+--
+-- Source evidence (each pair confirmed from migration files):
+--   houses_kadastr_idx2        (050:93) dup of houses_cadastral_number_idx (029:180)
+--   houses_addr_fp_idx         (028:20) dup of houses_fingerprint_idx2     (050:90)
+--   listings_scraped_idx       (002:59) dup of listings_scraped_desc_idx   (050:62)
+--   listing_sources_price_divergence_idx (042:8) dup of _idx2              (050:84)
+--   houses_source_ext_idx      (009:59) redundant — covered by UNIQUE(source,ext_house_id) (009:54)
+--
+-- Dependencies:
+--   002_core_tables.sql, 009_houses.sql, 028_matching_tables.sql,
+--   029_extend_matching_valuation_dynamics.sql, 042_listing_sources_price_divergence_idx.sql,
+--   050_search_optimization.sql
+
+BEGIN;
+
+-- =============================================================================
+-- Step 1: Drop exact-duplicate and redundant indexes
+-- =============================================================================
+
+-- dup of houses_cadastral_number_idx (029_extend_matching_valuation_dynamics.sql:180)
+-- same: ON houses (cadastral_number) WHERE cadastral_number IS NOT NULL
+DROP INDEX IF EXISTS houses_kadastr_idx2;
+
+-- dup of houses_fingerprint_idx2 (050_search_optimization.sql:90)
+-- same: ON houses (address_fingerprint) WHERE address_fingerprint IS NOT NULL
+DROP INDEX IF EXISTS houses_addr_fp_idx;
+
+-- dup of listings_scraped_desc_idx (050_search_optimization.sql:62)
+-- same: ON listings (scraped_at DESC) WHERE is_active = true
+-- Note: listings_scraped_desc_idx is also de-partialed in Step 2 below.
+DROP INDEX IF EXISTS listings_scraped_idx;
+
+-- dup of listing_sources_price_divergence_idx2 (050_search_optimization.sql:84)
+-- same: ON listing_sources (listing_id, price_rub) WHERE price_rub IS NOT NULL
+DROP INDEX IF EXISTS listing_sources_price_divergence_idx;
+
+-- redundant: covered by UNIQUE (source, ext_house_id) constraint (009_houses.sql:54)
+-- which automatically creates a unique index on the same column set.
+-- The plain btree here (009_houses.sql:59) is a strict duplicate of that unique index.
+DROP INDEX IF EXISTS houses_source_ext_idx;
+
+-- =============================================================================
+-- Step 2: De-partial is_active=true indexes
+-- Since is_active is 100% true, the WHERE predicate saves nothing.
+-- A full index strictly covers the partial — no plan regression.
+-- =============================================================================
+
+-- listings_geom_active_idx (050_search_optimization.sql:47)
+-- Original: ON listings USING GIST (geom) WHERE is_active = true
+DROP INDEX IF EXISTS listings_geom_active_idx;
+CREATE INDEX IF NOT EXISTS listings_geom_active_idx
+    ON listings USING GIST (geom);
+
+-- listings_active_filter_idx (050_search_optimization.sql:52)
+-- Original: ON listings (rooms, price_rub, area_m2, scraped_at DESC) WHERE is_active = true
+DROP INDEX IF EXISTS listings_active_filter_idx;
+CREATE INDEX IF NOT EXISTS listings_active_filter_idx
+    ON listings (rooms, price_rub, area_m2, scraped_at DESC);
+
+-- listings_house_price_idx (050_search_optimization.sql:57)
+-- Original: ON listings (house_id_fk, price_rub) WHERE is_active = true
+DROP INDEX IF EXISTS listings_house_price_idx;
+CREATE INDEX IF NOT EXISTS listings_house_price_idx
+    ON listings (house_id_fk, price_rub);
+
+-- listings_scraped_desc_idx (050_search_optimization.sql:62)
+-- Original: ON listings (scraped_at DESC) WHERE is_active = true
+-- Note: listings_scraped_idx (exact duplicate) already dropped in Step 1 above.
+DROP INDEX IF EXISTS listings_scraped_desc_idx;
+CREATE INDEX IF NOT EXISTS listings_scraped_desc_idx
+    ON listings (scraped_at DESC);
+
+-- listings_rooms_area_idx (002_core_tables.sql:58)
+-- Original: ON listings (rooms, area_m2) WHERE is_active = true
+DROP INDEX IF EXISTS listings_rooms_area_idx;
+CREATE INDEX IF NOT EXISTS listings_rooms_area_idx
+    ON listings (rooms, area_m2);
+
+COMMIT;