diff --git a/migrations/2025-11-01-114838_actors/up.sql b/migrations/2025-11-01-114838_actors/up.sql
index d08e17ad..2061e3e4 100644
--- a/migrations/2025-11-01-114838_actors/up.sql
+++ b/migrations/2025-11-01-114838_actors/up.sql
@@ -13,7 +13,6 @@
 
 CREATE TYPE actor_status AS ENUM ('active', 'takendown', 'suspended', 'deleted', 'deactivated');
 CREATE TYPE actor_sync_state AS ENUM ('synced', 'dirty', 'partial', 'processing');
-CREATE TYPE unresolved_reason AS ENUM ('pending', 'deleted', 'forbidden', 'not_found');
 
 -- =============================================================================
 -- PREFERENCE TYPES
@@ -40,41 +39,16 @@ CREATE TABLE actors (
     sync_state actor_sync_state NOT NULL DEFAULT 'dirty',
     repo_rev TEXT,
     repo_cid BYTEA,
-    last_indexed TIMESTAMPTZ
+    last_indexed TIMESTAMPTZ,
+    account_created_at TIMESTAMPTZ
 );
 
 CREATE INDEX idx_actors_handle ON actors(handle) WHERE handle IS NOT NULL;
 CREATE INDEX idx_actors_sync_state ON actors(sync_state);
 CREATE INDEX idx_actors_status ON actors(status);
 CREATE INDEX idx_actors_last_indexed ON actors(last_indexed) WHERE last_indexed IS NOT NULL;
+CREATE INDEX idx_actors_account_created_at ON actors(account_created_at) WHERE account_created_at IS NOT NULL;
 
--- =============================================================================
--- UNRESOLVED REFERENCES
--- =============================================================================
--- Tracks posts that haven't been received yet or have been deleted.
--- Enables 100 replies to reference a single unresolved CID via FK reuse.
--- Stores AT URI components (actor_did, rkey) for reconstruction.
-
-CREATE TABLE unresolved_refs (
-    id SERIAL PRIMARY KEY,
-    cid BYTEA NOT NULL UNIQUE,
-    reason unresolved_reason NOT NULL,
-    reference_count INTEGER NOT NULL DEFAULT 0,
-    actor_did TEXT NOT NULL,
-    rkey INT8 NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    last_checked TIMESTAMPTZ
-);
-
-CREATE INDEX idx_unresolved_refs_reason ON unresolved_refs(reason);
-CREATE INDEX idx_unresolved_refs_created_at ON unresolved_refs(created_at);
-
--- Allowlist for managing which actors to index
-CREATE TABLE allowlist (
-    actor_id INTEGER PRIMARY KEY REFERENCES actors(id) ON DELETE CASCADE,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    description TEXT
-);
 
 -- =============================================================================
 -- PROFILES
diff --git a/migrations/2025-11-01-114839_posts/down.sql b/migrations/2025-11-01-114839_posts/down.sql
index 01c7ed69..bcf568b5 100644
--- a/migrations/2025-11-01-114839_posts/down.sql
+++ b/migrations/2025-11-01-114839_posts/down.sql
@@ -13,9 +13,15 @@ DROP TABLE IF EXISTS post_embed_images;
 DROP TABLE IF EXISTS posts;
 DROP TABLE IF EXISTS uris;
 
+-- Drop functions
+DROP FUNCTION IF EXISTS i64_to_tid(BIGINT);
+DROP FUNCTION IF EXISTS tid_to_i64(TEXT);
+DROP FUNCTION IF EXISTS tid_timestamp(BIGINT);
+
 -- Drop types
 DROP TYPE IF EXISTS facet_type;
 DROP TYPE IF EXISTS language_code;
 DROP TYPE IF EXISTS caption_mime_type;
 DROP TYPE IF EXISTS video_mime_type;
 DROP TYPE IF EXISTS embed_type;
+DROP TYPE IF EXISTS post_status;
diff --git a/migrations/2025-11-01-114839_posts/up.sql b/migrations/2025-11-01-114839_posts/up.sql
index b29ccb87..4a0a73b4 100644
--- a/migrations/2025-11-01-114839_posts/up.sql
+++ b/migrations/2025-11-01-114839_posts/up.sql
@@ -11,6 +11,7 @@
 -- POST & EMBED TYPES
 -- =============================================================================
 
+CREATE TYPE post_status AS ENUM ('complete', 'stub', 'deleted', 'forbidden');
 CREATE TYPE embed_type AS ENUM ('images', 'video', 'external', 'record', 'record_with_media');
 CREATE TYPE video_mime_type AS ENUM ('video/mp4', 'video/webm', 'video/quicktime', 'video/mpeg');
 CREATE TYPE caption_mime_type AS ENUM ('text/vtt');
@@ -47,46 +48,102 @@ CREATE TYPE facet_type AS ENUM ('mention', 'link', 'tag');
 -- POSTS
 -- =============================================================================
 
+-- =============================================================================
+-- TID HELPER FUNCTIONS
+-- =============================================================================
+
+-- Extract timestamp from TID (INT8)
+CREATE OR REPLACE FUNCTION tid_timestamp(tid_int8 BIGINT)
+RETURNS TIMESTAMPTZ AS $$
+BEGIN
+    -- Extract timestamp (53 bits) by right shifting 10 bits (clock ID)
+    -- Convert microseconds to PostgreSQL timestamp
+    RETURN to_timestamp((tid_int8 >> 10) / 1000000.0);
+END;
+$$ LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
+
+-- Decode base32-sortable TID string to INT8
+CREATE OR REPLACE FUNCTION tid_to_i64(tid TEXT) RETURNS BIGINT AS $$
+DECLARE
+    alphabet TEXT := '234567abcdefghijklmnopqrstuvwxyz';
+    result BIGINT := 0;
+    char_pos INTEGER;
+    value INTEGER;
+    i INTEGER;
+BEGIN
+    IF LENGTH(tid) != 13 THEN
+        RAISE EXCEPTION 'TID must be 13 characters, got %', LENGTH(tid);
+    END IF;
+    IF POSITION(SUBSTRING(tid FROM 1 FOR 1) IN '234567abcdefghij') = 0 THEN
+        RAISE EXCEPTION 'Invalid first character in TID: %', SUBSTRING(tid FROM 1 FOR 1);
+    END IF;
+    FOR i IN 1..13 LOOP
+        char_pos := POSITION(SUBSTRING(tid FROM i FOR 1) IN alphabet);
+        IF char_pos = 0 THEN
+            RAISE EXCEPTION 'Invalid character in TID at position %: %', i, SUBSTRING(tid FROM i FOR 1);
+        END IF;
+        value := char_pos - 1;
+        result := (result << 5) | value;
+    END LOOP;
+    RETURN result;
+END;
+$$ LANGUAGE plpgsql IMMUTABLE STRICT;
+
+COMMENT ON FUNCTION tid_to_i64(TEXT) IS 'Decodes a base32-sortable TID string (13 characters) to a BIGINT. Uses AT Protocol TID encoding: https://atproto.com/specs/tid';
+
+-- Encode INT8 to base32-sortable TID string
+CREATE OR REPLACE FUNCTION i64_to_tid(value BIGINT) RETURNS TEXT AS $$
+DECLARE
+    alphabet TEXT := '234567abcdefghijklmnopqrstuvwxyz';
+    result TEXT := '';
+    char_index INTEGER;
+    i INTEGER;
+BEGIN
+    FOR i IN 1..13 LOOP
+        char_index := (value & 31) + 1;
+        result := SUBSTRING(alphabet FROM char_index FOR 1) || result;
+        value := value >> 5;
+    END LOOP;
+    RETURN result;
+END;
+$$ LANGUAGE plpgsql IMMUTABLE STRICT;
+
+COMMENT ON FUNCTION i64_to_tid(BIGINT) IS 'Encodes a BIGINT to a base32-sortable TID string (13 characters). Uses AT Protocol TID encoding: https://atproto.com/specs/tid';
+
 CREATE TABLE posts (
     id BIGSERIAL PRIMARY KEY,
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    content TEXT NOT NULL,
+    content BYTEA,
+    content_version SMALLINT,
     language_code language_code,
     tags TEXT[] NOT NULL DEFAULT '{}',
-    -- Resolved references (post exists in our database)
     parent_post_id BIGINT REFERENCES posts(id) ON DELETE SET NULL,
     root_post_id BIGINT REFERENCES posts(id) ON DELETE SET NULL,
-    -- Unresolved references (post pending/deleted/forbidden/not_found)
-    unresolved_parent_id INTEGER REFERENCES unresolved_refs(id) ON DELETE SET NULL,
-    unresolved_root_id INTEGER REFERENCES unresolved_refs(id) ON DELETE SET NULL,
     embed_type embed_type,
     embed_subtype embed_type,
     violates_threadgate BOOLEAN NOT NULL DEFAULT FALSE,
+    status post_status NOT NULL DEFAULT 'complete',
+    search_vector tsvector,
     UNIQUE (actor_id, rkey),
-    -- Ensure exactly one parent reference type (resolved XOR unresolved, or neither for root posts)
-    CHECK (
-        (parent_post_id IS NULL AND unresolved_parent_id IS NULL) OR  -- Root post
-        (parent_post_id IS NOT NULL AND unresolved_parent_id IS NULL) OR  -- Resolved
-        (parent_post_id IS NULL AND unresolved_parent_id IS NOT NULL)     -- Unresolved
-    ),
-    -- Ensure exactly one root reference type
     CHECK (
-        (root_post_id IS NULL AND unresolved_root_id IS NULL) OR
-        (root_post_id IS NOT NULL AND unresolved_root_id IS NULL) OR
-        (root_post_id IS NULL AND unresolved_root_id IS NOT NULL)
+        (content IS NULL AND content_version IS NULL) OR
+        (content IS NOT NULL AND content_version IS NOT NULL)
     )
 );
 
+COMMENT ON COLUMN posts.content IS 'Zstd-compressed post content (BYTEA). NULL for stub/deleted posts.';
+COMMENT ON COLUMN posts.content_version IS 'Dictionary version (NULL if no content)';
+COMMENT ON COLUMN posts.search_vector IS 'Full-text search vector (NULL except for allowlisted users)';
+COMMENT ON COLUMN posts.status IS 'Post lifecycle state: complete (normal), stub (placeholder), deleted (soft-deleted), forbidden (access denied)';
+
 CREATE INDEX idx_posts_actor ON posts(actor_id);
-CREATE INDEX idx_posts_created_at ON posts(created_at DESC);
 CREATE INDEX idx_posts_parent ON posts(parent_post_id) WHERE parent_post_id IS NOT NULL;
 CREATE INDEX idx_posts_root ON posts(root_post_id) WHERE root_post_id IS NOT NULL;
-CREATE INDEX idx_posts_unresolved_parent ON posts(unresolved_parent_id) WHERE unresolved_parent_id IS NOT NULL;
-CREATE INDEX idx_posts_unresolved_root ON posts(unresolved_root_id) WHERE unresolved_root_id IS NOT NULL;
 CREATE INDEX idx_posts_tags ON posts USING GIN(tags);
+CREATE INDEX idx_posts_status ON posts(status) WHERE status != 'complete';
+CREATE INDEX idx_posts_search_vector ON posts USING GIN (search_vector) WHERE search_vector IS NOT NULL;
 
 -- Add foreign key constraints to profiles and statuses
 ALTER TABLE profiles ADD CONSTRAINT fk_profiles_pinned_post
diff --git a/migrations/2025-11-01-114840_social_graph/up.sql b/migrations/2025-11-01-114840_social_graph/up.sql
index d2806591..0b8f7adc 100644
--- a/migrations/2025-11-01-114840_social_graph/up.sql
+++ b/migrations/2025-11-01-114840_social_graph/up.sql
@@ -12,7 +12,6 @@ CREATE TABLE follows (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     subject_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     UNIQUE (actor_id, rkey),
     UNIQUE (actor_id, subject_actor_id)
@@ -27,7 +26,6 @@ CREATE TABLE blocks (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     subject_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     UNIQUE (actor_id, rkey),
     UNIQUE (actor_id, subject_actor_id)
diff --git a/migrations/2025-11-01-114841_engagement/down.sql b/migrations/2025-11-01-114841_engagement/down.sql
index be391803..284b2b8a 100644
--- a/migrations/2025-11-01-114841_engagement/down.sql
+++ b/migrations/2025-11-01-114841_engagement/down.sql
@@ -2,3 +2,4 @@ DROP TABLE IF EXISTS bookmarks;
 DROP TABLE IF EXISTS reposts;
 DROP TABLE IF EXISTS likes;
 DROP TYPE IF EXISTS like_subject_type;
+DROP TYPE IF EXISTS bookmark_subject_type;
diff --git a/migrations/2025-11-01-114841_engagement/up.sql b/migrations/2025-11-01-114841_engagement/up.sql
index b1d616bf..e7893518 100644
--- a/migrations/2025-11-01-114841_engagement/up.sql
+++ b/migrations/2025-11-01-114841_engagement/up.sql
@@ -11,6 +11,7 @@
 -- =============================================================================
 
 CREATE TYPE like_subject_type AS ENUM ('post', 'feedgen', 'labeler');
+CREATE TYPE bookmark_subject_type AS ENUM ('post', 'feedgen', 'list', 'labeler', 'starterpack');
 
 -- =============================================================================
 -- LIKES
@@ -22,20 +23,14 @@ CREATE TABLE likes (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     subject_type like_subject_type NOT NULL DEFAULT 'post',
-    subject_id BIGINT, -- Nullable for pending linkage
+    subject_id BIGINT,
     via_post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE,
-    -- Pending linkage for likes that arrive before their subjects
-    pending_subject_cid BYTEA,
-    pending_via_post_cid BYTEA,
     UNIQUE (actor_id, rkey)
 );
 
 CREATE INDEX idx_likes_actor ON likes(actor_id);
 CREATE INDEX idx_likes_subject ON likes(subject_type, subject_id);
-CREATE INDEX idx_likes_pending_subject_cid ON likes(pending_subject_cid) WHERE pending_subject_cid IS NOT NULL;
-CREATE INDEX idx_likes_pending_via_post_cid ON likes(pending_via_post_cid) WHERE pending_via_post_cid IS NOT NULL;
 
 -- =============================================================================
 -- REPOSTS
@@ -46,31 +41,25 @@ CREATE TABLE reposts (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE, -- Nullable for pending linkage
+    post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE,
     via_post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE,
-    -- Pending linkage for reposts that arrive before their targets
-    pending_post_cid BYTEA,
-    pending_via_post_cid BYTEA,
-    UNIQUE (actor_id, rkey),
-    UNIQUE (actor_id, post_id)
+    UNIQUE (actor_id, rkey)
 );
 
 CREATE INDEX idx_reposts_actor ON reposts(actor_id);
 CREATE INDEX idx_reposts_post ON reposts(post_id);
-CREATE INDEX idx_reposts_pending_post_cid ON reposts(pending_post_cid) WHERE pending_post_cid IS NOT NULL;
-CREATE INDEX idx_reposts_pending_via_post_cid ON reposts(pending_via_post_cid) WHERE pending_via_post_cid IS NOT NULL;
 
 -- =============================================================================
 -- BOOKMARKS
 -- =============================================================================
 
--- Bookmarks (off-protocol, almost always posts)
+-- Bookmarks (off-protocol, supports multiple entity types)
 CREATE TABLE bookmarks (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
-    subject_post_id BIGINT NOT NULL REFERENCES posts(id) ON DELETE CASCADE,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    PRIMARY KEY (actor_id, subject_post_id)
+    rkey INT8 NOT NULL,
+    subject_type bookmark_subject_type NOT NULL,
+    subject_id BIGINT,
+    PRIMARY KEY (actor_id, rkey)
 );
 
-CREATE INDEX idx_bookmarks_actor_created ON bookmarks(actor_id, created_at DESC);
+CREATE INDEX idx_bookmarks_subject ON bookmarks(subject_type, subject_id) WHERE subject_id IS NOT NULL;
diff --git a/migrations/2025-11-01-114842_notifications/up.sql b/migrations/2025-11-01-114842_notifications/up.sql
index 4a402ddd..488a173c 100644
--- a/migrations/2025-11-01-114842_notifications/up.sql
+++ b/migrations/2025-11-01-114842_notifications/up.sql
@@ -1,43 +1,13 @@
 -- =============================================================================
--- NOTIFICATIONS
+-- NOTIFICATION PREFERENCES
 -- =============================================================================
 --
--- Notification system: notifications, seen tracking, preferences
+-- Notification preferences only (notifications are stored in Redis)
 --
 -- =============================================================================
 
--- =============================================================================
--- NOTIFICATION TYPES
--- =============================================================================
-
-CREATE TYPE notification_record_type AS ENUM ('post', 'like', 'repost', 'follow', 'block');
-CREATE TYPE notification_reason AS ENUM ('follow', 'like', 'repost', 'reply', 'mention', 'quote');
 CREATE TYPE notif_allow_subscriptions AS ENUM ('none', 'mutuals', 'followers');
 
--- =============================================================================
--- NOTIFICATIONS
--- =============================================================================
-
-CREATE TABLE notifications (
-    recipient_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
-    record_type notification_record_type NOT NULL,
-    record_id BIGINT NOT NULL,
-    reason notification_reason NOT NULL,
-    reason_subject_post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE,
-    is_read BOOLEAN NOT NULL DEFAULT FALSE,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    PRIMARY KEY (recipient_actor_id, record_type, record_id)
-);
-
-CREATE INDEX idx_notifications_recipient_created ON notifications(recipient_actor_id, created_at DESC);
-CREATE INDEX idx_notifications_recipient_unread ON notifications(recipient_actor_id) WHERE is_read = FALSE;
-
--- Notification seen tracking
-CREATE TABLE notification_seens (
-    actor_id INTEGER PRIMARY KEY REFERENCES actors(id) ON DELETE CASCADE,
-    seen_at TIMESTAMPTZ NOT NULL
-);
-
 -- Notification preferences
 CREATE TABLE notif_decl (
     actor_id INTEGER PRIMARY KEY REFERENCES actors(id) ON DELETE CASCADE,
diff --git a/migrations/2025-11-01-114843_thread_moderation/down.sql b/migrations/2025-11-01-114843_thread_moderation/down.sql
index 382216e9..2fe0e6f7 100644
--- a/migrations/2025-11-01-114843_thread_moderation/down.sql
+++ b/migrations/2025-11-01-114843_thread_moderation/down.sql
@@ -4,5 +4,6 @@ DROP TABLE IF EXISTS postgates;
 DROP TABLE IF EXISTS threadgate_allowed_lists;
 DROP TABLE IF EXISTS threadgate_hidden_replies;
 DROP TABLE IF EXISTS threadgates;
+DROP FUNCTION IF EXISTS maintain_postgates(TEXT, TEXT[], TIMESTAMP);
 DROP TYPE IF EXISTS postgate_rule;
 DROP TYPE IF EXISTS threadgate_rule;
diff --git a/migrations/2025-11-01-114843_thread_moderation/up.sql b/migrations/2025-11-01-114843_thread_moderation/up.sql
index 4877d9e3..e038714d 100644
--- a/migrations/2025-11-01-114843_thread_moderation/up.sql
+++ b/migrations/2025-11-01-114843_thread_moderation/up.sql
@@ -32,16 +32,12 @@ CREATE TABLE threadgates (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE, -- Nullable for pending linkage
+    post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE,
     allow threadgate_rule[],
-    -- Pending linkage for threadgates that arrive before their posts
-    pending_post_cid BYTEA,
     UNIQUE (actor_id, rkey)
 );
 
 CREATE INDEX idx_threadgates_post ON threadgates(post_id);
-CREATE INDEX idx_threadgates_pending_post_cid ON threadgates(pending_post_cid) WHERE pending_post_cid IS NOT NULL;
 
 -- Threadgate Hidden Replies
 CREATE TABLE threadgate_hidden_replies (
@@ -66,16 +62,12 @@ CREATE TABLE postgates (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE, -- Nullable for pending linkage
+    post_id BIGINT REFERENCES posts(id) ON DELETE CASCADE,
     rules postgate_rule[] NOT NULL DEFAULT '{}',
-    -- Pending linkage for postgates that arrive before their posts
-    pending_post_cid BYTEA,
     UNIQUE (actor_id, rkey)
 );
 
 CREATE INDEX idx_postgates_post ON postgates(post_id);
-CREATE INDEX idx_postgates_pending_post_cid ON postgates(pending_post_cid) WHERE pending_post_cid IS NOT NULL;
 
 -- Postgate Detached Embeds
 CREATE TABLE postgate_detached (
@@ -84,6 +76,66 @@ CREATE TABLE postgate_detached (
     PRIMARY KEY (postgate_id, detached_post_id)
 );
 
+-- =============================================================================
+-- POSTGATE MAINTENANCE FUNCTION
+-- =============================================================================
+
+-- Updates the 'detached' flag on existing post_embed_record entries when a postgate changes
+CREATE OR REPLACE FUNCTION maintain_postgates(
+    post_uri TEXT,
+    detached_uris TEXT[],
+    disable_effective TIMESTAMP
+) RETURNS BIGINT AS $$
+DECLARE
+    target_post_id BIGINT;
+    rows_updated BIGINT;
+BEGIN
+    -- Look up the post_id from the post URI
+    SELECT p.id INTO target_post_id
+    FROM posts p
+    INNER JOIN actors a ON p.actor_id = a.id
+    WHERE 'at://' || a.did || '/app.bsky.feed.post/' || i64_to_tid(p.rkey) = post_uri;
+
+    IF target_post_id IS NULL THEN
+        RETURN 0;
+    END IF;
+
+    -- Update post_embed_record entries to mark quotes as detached
+    WITH detached_post_ids AS (
+        SELECT p.id
+        FROM UNNEST(detached_uris) AS uri
+        CROSS JOIN LATERAL (
+            SELECT
+                SPLIT_PART(SUBSTRING(uri FROM 6), '/', 1) as did,
+                SPLIT_PART(SUBSTRING(uri FROM 6), '/', 3) as rkey
+        ) parts
+        INNER JOIN actors a ON a.did = parts.did
+        INNER JOIN posts p ON p.actor_id = a.id AND p.rkey = tid_to_i64(parts.rkey)
+    ),
+    posts_to_update AS (
+        SELECT per.post_id
+        FROM post_embed_record per
+        WHERE per.embedded_post_id = target_post_id
+    )
+    UPDATE post_embed_record per
+    SET detached = (
+        per.post_id IN (SELECT id FROM detached_post_ids)
+        OR
+        (disable_effective IS NOT NULL AND (
+            SELECT tid_timestamp(p.rkey)
+            FROM posts p
+            WHERE p.id = per.post_id
+        ) > disable_effective)
+    )
+    WHERE per.embedded_post_id = target_post_id;
+
+    GET DIAGNOSTICS rows_updated = ROW_COUNT;
+    RETURN rows_updated;
+END;
+$$ LANGUAGE plpgsql;
+
+COMMENT ON FUNCTION maintain_postgates IS 'Updates detached flag on post_embed_record when postgates change';
+
 -- =============================================================================
 -- THREAD MUTES
 -- =============================================================================
diff --git a/migrations/2025-11-01-114844_lists/down.sql b/migrations/2025-11-01-114844_lists/down.sql
index 05c47737..b43cf8d6 100644
--- a/migrations/2025-11-01-114844_lists/down.sql
+++ b/migrations/2025-11-01-114844_lists/down.sql
@@ -4,3 +4,4 @@ DROP TABLE IF EXISTS list_blocks;
 DROP TABLE IF EXISTS list_items;
 DROP TABLE IF EXISTS lists;
 DROP TYPE IF EXISTS list_type;
+DROP TYPE IF EXISTS record_status;
diff --git a/migrations/2025-11-01-114844_lists/up.sql b/migrations/2025-11-01-114844_lists/up.sql
index 794a725c..24f5acea 100644
--- a/migrations/2025-11-01-114844_lists/up.sql
+++ b/migrations/2025-11-01-114844_lists/up.sql
@@ -16,6 +16,8 @@ CREATE TYPE list_type AS ENUM (
     'app.bsky.graph.defs#referencelist'
 );
 
+CREATE TYPE record_status AS ENUM ('complete', 'stub', 'deleted', 'forbidden');
+
 -- =============================================================================
 -- LISTS
 -- =============================================================================
@@ -25,18 +27,22 @@ CREATE TABLE lists (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     owner_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
-    list_type list_type NOT NULL,
-    name TEXT NOT NULL,
+    list_type list_type,
+    name TEXT,
     description TEXT,
     description_facets JSONB,
     avatar_cid BYTEA,
+    status record_status NOT NULL DEFAULT 'complete',
     UNIQUE (actor_id, rkey)
 );
 
+COMMENT ON COLUMN lists.status IS 'Status of the list record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = user deleted, ''forbidden'' = removed by moderators.';
+COMMENT ON COLUMN lists.list_type IS 'Type of list (e.g., modlist, curatelist). NULL for stub lists (not yet fetched).';
+COMMENT ON COLUMN lists.name IS 'List display name. NULL for stub lists (not yet fetched).';
+
 CREATE INDEX idx_lists_owner ON lists(owner_actor_id);
-CREATE INDEX idx_lists_type ON lists(list_type);
+CREATE INDEX idx_lists_type ON lists(list_type) WHERE list_type IS NOT NULL;
 
 -- =============================================================================
 -- LIST ITEMS
@@ -47,17 +53,13 @@ CREATE TABLE list_items (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    list_id BIGINT REFERENCES lists(id) ON DELETE CASCADE, -- Nullable for pending linkage
+    list_id BIGINT NOT NULL REFERENCES lists(id) ON DELETE CASCADE,
     subject_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
-    -- Pending linkage for list items that arrive before their lists
-    pending_list_cid BYTEA,
     UNIQUE (actor_id, rkey)
 );
 
 CREATE INDEX idx_list_items_list ON list_items(list_id);
 CREATE INDEX idx_list_items_subject ON list_items(subject_actor_id);
-CREATE INDEX idx_list_items_pending_list_cid ON list_items(pending_list_cid) WHERE pending_list_cid IS NOT NULL;
 
 -- =============================================================================
 -- LIST BLOCKS & MUTES
@@ -69,7 +71,6 @@ CREATE TABLE list_blocks (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     list_id BIGINT NOT NULL REFERENCES lists(id) ON DELETE CASCADE,
     UNIQUE (actor_id, rkey)
 );
diff --git a/migrations/2025-11-01-114845_feeds/down.sql b/migrations/2025-11-01-114845_feeds/down.sql
index 44ebe6ea..21b8d15d 100644
--- a/migrations/2025-11-01-114845_feeds/down.sql
+++ b/migrations/2025-11-01-114845_feeds/down.sql
@@ -3,3 +3,5 @@ DROP TABLE IF EXISTS starterpack_feeds;
 DROP TABLE IF EXISTS starterpacks;
 DROP TABLE IF EXISTS feedgens;
 DROP TYPE IF EXISTS content_mode;
+DROP TYPE IF EXISTS feedgen_status;
+DROP TYPE IF EXISTS labeler_status;
diff --git a/migrations/2025-11-01-114845_feeds/up.sql b/migrations/2025-11-01-114845_feeds/up.sql
index 71479d1f..f3140aac 100644
--- a/migrations/2025-11-01-114845_feeds/up.sql
+++ b/migrations/2025-11-01-114845_feeds/up.sql
@@ -11,6 +11,8 @@
 -- =============================================================================
 
 CREATE TYPE content_mode AS ENUM ('contentModeUnspecified', 'contentModeVideo');
+CREATE TYPE feedgen_status AS ENUM ('complete', 'stub', 'deleted');
+CREATE TYPE labeler_status AS ENUM ('complete', 'stub', 'deleted');
 
 -- =============================================================================
 -- FEED GENERATORS
@@ -19,20 +21,25 @@ CREATE TYPE content_mode AS ENUM ('contentModeUnspecified', 'contentModeVideo');
 CREATE TABLE feedgens (
     id BIGSERIAL PRIMARY KEY,
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
-    rkey INT8 NOT NULL,
+    rkey TEXT NOT NULL,
     cid BYTEA NOT NULL,
     created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     owner_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     service_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     content_mode content_mode,
-    name TEXT NOT NULL,
+    name TEXT,
     description TEXT,
     description_facets JSONB,
     avatar_cid BYTEA,
     accepts_interactions BOOLEAN NOT NULL DEFAULT FALSE,
+    status feedgen_status NOT NULL DEFAULT 'complete',
     UNIQUE (actor_id, rkey)
 );
 
+COMMENT ON COLUMN feedgens.rkey IS 'Feedgen record key. Can be arbitrary string (e.g., "music", "discover") or TID format.';
+COMMENT ON COLUMN feedgens.status IS 'Status of the feedgen record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = soft-deleted.';
+COMMENT ON COLUMN feedgens.name IS 'Feedgen display name. NULL for stub feedgens (not yet fetched) or deleted feedgens (soft-deleted).';
+
 CREATE INDEX idx_feedgens_owner ON feedgens(owner_actor_id);
 CREATE INDEX idx_feedgens_service ON feedgens(service_actor_id);
 
@@ -45,26 +52,25 @@ CREATE TABLE starterpacks (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     owner_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
-    name TEXT NOT NULL,
+    name TEXT,
     description TEXT,
     description_facets JSONB,
     list_id BIGINT NOT NULL REFERENCES lists(id) ON DELETE CASCADE,
+    search_vector tsvector,
+    status record_status NOT NULL DEFAULT 'complete',
     UNIQUE (actor_id, rkey)
 );
 
+COMMENT ON COLUMN starterpacks.status IS 'Status of the starter pack record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = user deleted, ''forbidden'' = removed by moderators.';
+COMMENT ON COLUMN starterpacks.name IS 'Starter pack name. NULL for stub starter packs (not yet fetched).';
+
 CREATE INDEX idx_starterpacks_owner ON starterpacks(owner_actor_id);
 CREATE INDEX idx_starterpacks_list ON starterpacks(list_id);
 
 -- Full-text search on starter packs
-ALTER TABLE starterpacks ADD COLUMN search_vector tsvector
-    GENERATED ALWAYS AS (
-        setweight(to_tsvector('english', coalesce(name, '')), 'A') ||
-        setweight(to_tsvector('english', coalesce(description, '')), 'B')
-    ) STORED;
-
-CREATE INDEX idx_starterpacks_search ON starterpacks USING GIN(search_vector);
+-- Using a manual update approach instead of GENERATED ALWAYS AS since name can be NULL
+CREATE INDEX idx_starterpacks_search ON starterpacks USING GIN(search_vector) WHERE search_vector IS NOT NULL;
 
 -- Starter Pack Feeds
 CREATE TABLE starterpack_feeds (
diff --git a/migrations/2025-11-01-114846_moderation/up.sql b/migrations/2025-11-01-114846_moderation/up.sql
index a35180a0..800f3037 100644
--- a/migrations/2025-11-01-114846_moderation/up.sql
+++ b/migrations/2025-11-01-114846_moderation/up.sql
@@ -26,9 +26,12 @@ CREATE TABLE labelers (
     created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     reasons reason_type[],
     subject_types subject_type[],
-    subject_collections TEXT[]
+    subject_collections TEXT[],
+    status labeler_status NOT NULL DEFAULT 'complete'
 );
 
+COMMENT ON COLUMN labelers.status IS 'Status of the labeler record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = soft-deleted.';
+
 -- Labeler Definitions
 CREATE TABLE labeler_defs (
     labeler_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
@@ -72,7 +75,6 @@ CREATE TABLE verification (
     actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     rkey INT8 NOT NULL,
     cid BYTEA NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
     verifier_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     subject_actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
     handle TEXT NOT NULL,
diff --git a/migrations/2025-11-01-192943_bookmarks_discriminated_union/down.sql b/migrations/2025-11-01-192943_bookmarks_discriminated_union/down.sql
deleted file mode 100644
index abfd36a7..00000000
--- a/migrations/2025-11-01-192943_bookmarks_discriminated_union/down.sql
+++ /dev/null
@@ -1,48 +0,0 @@
--- =============================================================================
--- BOOKMARKS: Revert Discriminated Union Pattern
--- =============================================================================
---
--- Rollback migration: revert bookmarks table to post-only schema
---
--- WARNING: This migration will DROP all bookmarks for non-post subjects
--- (feedgens, lists, labelers, starter packs)
--- =============================================================================
-
--- Add back the old column
-ALTER TABLE bookmarks ADD COLUMN subject_post_id BIGINT;
-
--- Copy post bookmarks back
-UPDATE bookmarks
-SET subject_post_id = subject_id
-WHERE subject_type = 'post';
-
--- Drop new indexes
-DROP INDEX idx_bookmarks_subject;
-DROP INDEX idx_bookmarks_actor_created;
-
--- Drop new primary key
-ALTER TABLE bookmarks DROP CONSTRAINT bookmarks_pkey;
-
--- Delete non-post bookmarks (can't be preserved in old schema)
-DELETE FROM bookmarks WHERE subject_type != 'post';
-
--- Make subject_post_id NOT NULL
-ALTER TABLE bookmarks ALTER COLUMN subject_post_id SET NOT NULL;
-
--- Drop new columns
-ALTER TABLE bookmarks DROP COLUMN rkey;
-ALTER TABLE bookmarks DROP COLUMN subject_type;
-ALTER TABLE bookmarks DROP COLUMN subject_id;
-
--- Add back old primary key
-ALTER TABLE bookmarks ADD PRIMARY KEY (actor_id, subject_post_id);
-
--- Add back old foreign key
-ALTER TABLE bookmarks ADD CONSTRAINT bookmarks_subject_post_id_fkey
-    FOREIGN KEY (subject_post_id) REFERENCES posts(id) ON DELETE CASCADE;
-
--- Recreate old index
-CREATE INDEX idx_bookmarks_actor_created ON bookmarks(actor_id, created_at DESC);
-
--- Drop enum type
-DROP TYPE bookmark_subject_type;
diff --git a/migrations/2025-11-01-192943_bookmarks_discriminated_union/up.sql b/migrations/2025-11-01-192943_bookmarks_discriminated_union/up.sql
deleted file mode 100644
index da3aa4f5..00000000
--- a/migrations/2025-11-01-192943_bookmarks_discriminated_union/up.sql
+++ /dev/null
@@ -1,33 +0,0 @@
--- =============================================================================
--- BOOKMARKS: Upgrade to Discriminated Union Pattern
--- =============================================================================
---
--- Upgrade bookmarks table from post-only to support multiple entity types:
--- posts, feed generators, lists, labelers, and starter packs
---
--- Migration strategy:
--- 1. Create new enum type for subject types
--- 2. Add new columns (rkey, subject_type, subject_id)
--- 3. Drop old table and create new one (fresh database approach)
--- 4. Create new primary key and indexes
--- =============================================================================
-
--- Drop old bookmarks table completely
-DROP TABLE IF EXISTS bookmarks CASCADE;
-
--- Create enum for bookmark subject types
-CREATE TYPE bookmark_subject_type AS ENUM ('post', 'feedgen', 'list', 'labeler', 'starterpack');
-
--- Create new bookmarks table with INT8 rkeys
-CREATE TABLE bookmarks (
-    actor_id INTEGER NOT NULL REFERENCES actors(id) ON DELETE CASCADE,
-    rkey INT8 NOT NULL,
-    subject_type bookmark_subject_type NOT NULL,
-    subject_id BIGINT, -- Nullable for pending linkage
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    PRIMARY KEY (actor_id, rkey)
-);
-
--- Create indexes
-CREATE INDEX idx_bookmarks_actor_created ON bookmarks(actor_id, created_at DESC);
-CREATE INDEX idx_bookmarks_subject ON bookmarks(subject_type, subject_id) WHERE subject_id IS NOT NULL;
diff --git a/migrations/2025-11-01-234535_drop_allowlist_table/down.sql b/migrations/2025-11-01-234535_drop_allowlist_table/down.sql
deleted file mode 100644
index 6c356529..00000000
--- a/migrations/2025-11-01-234535_drop_allowlist_table/down.sql
+++ /dev/null
@@ -1,18 +0,0 @@
--- Recreate allowlist table from actors.sync_state
---
--- This rollback migration recreates the allowlist table and populates it
--- from actors with sync_state IN ('synced', 'dirty', 'processing')
-
--- Step 1: Recreate allowlist table
-CREATE TABLE allowlist (
-    actor_id INTEGER PRIMARY KEY REFERENCES actors(id) ON DELETE CASCADE,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    description TEXT
-);
-
--- Step 2: Repopulate from actors table
--- Actors with sync_state IN ('synced', 'dirty', 'processing') are allowlisted
-INSERT INTO allowlist (actor_id, created_at, description)
-SELECT id, COALESCE(last_indexed, NOW()), NULL
-FROM actors
-WHERE sync_state IN ('synced'::actor_sync_state, 'dirty'::actor_sync_state, 'processing'::actor_sync_state);
diff --git a/migrations/2025-11-01-234535_drop_allowlist_table/up.sql b/migrations/2025-11-01-234535_drop_allowlist_table/up.sql
deleted file mode 100644
index dabe2b18..00000000
--- a/migrations/2025-11-01-234535_drop_allowlist_table/up.sql
+++ /dev/null
@@ -1,25 +0,0 @@
--- Drop allowlist table (redundant with actors.sync_state column)
---
--- Allowlist membership is now determined by sync_state:
--- - sync_state IN ('synced', 'dirty', 'processing') = fully allowed
--- - sync_state = 'partial' = not allowed (interacts with allowlisted users)
---
--- This migration:
--- 1. Ensures all currently allowlisted DIDs have correct sync_state
--- 2. Drops the allowlist table
--- 3. Removes any triggers/functions (if they existed)
-
--- Step 1: Ensure all allowlisted actors have correct sync_state
--- (This is idempotent and safe to run multiple times)
-UPDATE actors
-SET sync_state = 'dirty'::actor_sync_state
-WHERE id IN (SELECT actor_id FROM allowlist)
-  AND sync_state = 'partial'::actor_sync_state;
-
--- Step 2: Drop allowlist table
-DROP TABLE IF EXISTS allowlist;
-
--- Step 3: Drop NOTIFY function/trigger (if it existed)
--- Note: These were never created in migrations, but drop them just in case
-DROP TRIGGER IF EXISTS allowlist_changes_trigger ON actors;
-DROP FUNCTION IF EXISTS notify_allowlist_changes() CASCADE;
diff --git a/migrations/2025-11-02-193056_add_tid_to_i64_function/down.sql b/migrations/2025-11-02-193056_add_tid_to_i64_function/down.sql
deleted file mode 100644
index 41cb2189..00000000
--- a/migrations/2025-11-02-193056_add_tid_to_i64_function/down.sql
+++ /dev/null
@@ -1,2 +0,0 @@
--- Drop the tid_to_i64 function
-DROP FUNCTION IF EXISTS tid_to_i64(TEXT);
diff --git a/migrations/2025-11-02-193056_add_tid_to_i64_function/up.sql b/migrations/2025-11-02-193056_add_tid_to_i64_function/up.sql
deleted file mode 100644
index f0cf626e..00000000
--- a/migrations/2025-11-02-193056_add_tid_to_i64_function/up.sql
+++ /dev/null
@@ -1,49 +0,0 @@
--- Create PostgreSQL function to decode base32-sortable TID strings to INT8
--- This function implements the same logic as parakeet-db/src/tid_util.rs::decode_tid
--- TIDs are 13-character base32-sortable encoded 64-bit integers
--- Alphabet: 234567abcdefghijklmnopqrstuvwxyz (32 chars = 5 bits each)
--- Valid first characters: 234567abcdefghij (timestamp constraint)
-
-CREATE OR REPLACE FUNCTION tid_to_i64(tid TEXT) RETURNS BIGINT AS $$
-DECLARE
-    alphabet TEXT := '234567abcdefghijklmnopqrstuvwxyz';
-    result BIGINT := 0;
-    char_pos INTEGER;
-    value INTEGER;
-    i INTEGER;
-BEGIN
-    -- Validate length
-    IF LENGTH(tid) != 13 THEN
-        RAISE EXCEPTION 'TID must be 13 characters, got %', LENGTH(tid);
-    END IF;
-
-    -- Validate first character (must be in range for valid timestamps)
-    IF POSITION(SUBSTRING(tid FROM 1 FOR 1) IN '234567abcdefghij') = 0 THEN
-        RAISE EXCEPTION 'Invalid first character in TID: %', SUBSTRING(tid FROM 1 FOR 1);
-    END IF;
-
-    -- Decode base32-sortable string
-    FOR i IN 1..13 LOOP
-        -- Find position of character in alphabet (1-indexed)
-        char_pos := POSITION(SUBSTRING(tid FROM i FOR 1) IN alphabet);
-
-        IF char_pos = 0 THEN
-            RAISE EXCEPTION 'Invalid character in TID at position %: %', i, SUBSTRING(tid FROM i FOR 1);
-        END IF;
-
-        -- Convert to 0-indexed value
-        value := char_pos - 1;
-
-        -- Shift left by 5 bits and add new value
-        result := (result << 5) | value;
-    END LOOP;
-
-    RETURN result;
-END;
-$$ LANGUAGE plpgsql IMMUTABLE STRICT;
-
--- Add comment for documentation
-COMMENT ON FUNCTION tid_to_i64(TEXT) IS
-'Decodes a base32-sortable TID string (13 characters) to a BIGINT.
-Uses AT Protocol TID encoding: https://atproto.com/specs/tid
-Alphabet: 234567abcdefghijklmnopqrstuvwxyz';
diff --git a/migrations/2025-11-02-194138_revert_feedgens_rkey_to_text/down.sql b/migrations/2025-11-02-194138_revert_feedgens_rkey_to_text/down.sql
deleted file mode 100644
index aa544dd7..00000000
--- a/migrations/2025-11-02-194138_revert_feedgens_rkey_to_text/down.sql
+++ /dev/null
@@ -1,8 +0,0 @@
--- This file should undo anything in `up.sql`
---
--- NOTE: This down migration is NOT safe to run if you have feedgens with
--- arbitrary string rkeys (non-TID format). It will fail with a type conversion error.
---
--- Only run this if you're certain all feedgen rkeys are valid TIDs.
---
-ALTER TABLE feedgens ALTER COLUMN rkey TYPE INT8 USING tid_to_i64(rkey);
diff --git a/migrations/2025-11-02-194138_revert_feedgens_rkey_to_text/up.sql b/migrations/2025-11-02-194138_revert_feedgens_rkey_to_text/up.sql
deleted file mode 100644
index fc5caa82..00000000
--- a/migrations/2025-11-02-194138_revert_feedgens_rkey_to_text/up.sql
+++ /dev/null
@@ -1,17 +0,0 @@
--- =============================================================================
--- Revert feedgens.rkey from INT8 to TEXT
--- =============================================================================
---
--- REASON: Feedgens allow arbitrary user-chosen string rkeys (not TIDs)
--- Example feedgen rkeys: "music", "discover", "whats-hot", "for-you"
---
--- Analysis of parakeet_historical shows:
--- - 856 total feedgens
--- - 553 (64.6%) use arbitrary string rkeys
--- - Only 303 (35.4%) use TID format
---
--- Other collections (posts, likes, reposts, etc.) correctly use INT8 for TIDs.
--- Feedgens are the exception that MUST remain TEXT.
--- =============================================================================
-
-ALTER TABLE feedgens ALTER COLUMN rkey TYPE TEXT;
diff --git a/migrations/2025-11-02-201750_add_i64_to_tid_function/down.sql b/migrations/2025-11-02-201750_add_i64_to_tid_function/down.sql
deleted file mode 100644
index 1dbd90f7..00000000
--- a/migrations/2025-11-02-201750_add_i64_to_tid_function/down.sql
+++ /dev/null
@@ -1,2 +0,0 @@
--- Drop i64_to_tid function
-DROP FUNCTION IF EXISTS i64_to_tid(BIGINT);
diff --git a/migrations/2025-11-02-201750_add_i64_to_tid_function/up.sql b/migrations/2025-11-02-201750_add_i64_to_tid_function/up.sql
deleted file mode 100644
index f3693d8e..00000000
--- a/migrations/2025-11-02-201750_add_i64_to_tid_function/up.sql
+++ /dev/null
@@ -1,33 +0,0 @@
--- Create PostgreSQL function to encode INT8 to base32-sortable TID strings
--- This function implements the reverse of tid_to_i64
--- TIDs are 13-character base32-sortable encoded 64-bit integers
--- Alphabet: 234567abcdefghijklmnopqrstuvwxyz (32 chars = 5 bits each)
-
-CREATE OR REPLACE FUNCTION i64_to_tid(value BIGINT) RETURNS TEXT AS $$
-DECLARE
-    alphabet TEXT := '234567abcdefghijklmnopqrstuvwxyz';
-    result TEXT := '';
-    char_index INTEGER;
-    i INTEGER;
-BEGIN
-    -- Encode to base32-sortable string (13 characters)
-    FOR i IN 1..13 LOOP
-        -- Extract rightmost 5 bits
-        char_index := (value & 31) + 1;  -- 31 = 0b11111, +1 for 1-indexed substring
-
-        -- Prepend character to result (we build from right to left)
-        result := SUBSTRING(alphabet FROM char_index FOR 1) || result;
-
-        -- Shift right by 5 bits
-        value := value >> 5;
-    END LOOP;
-
-    RETURN result;
-END;
-$$ LANGUAGE plpgsql IMMUTABLE STRICT;
-
--- Add comment for documentation
-COMMENT ON FUNCTION i64_to_tid(BIGINT) IS
-'Encodes a BIGINT to a base32-sortable TID string (13 characters).
-Uses AT Protocol TID encoding: https://atproto.com/specs/tid
-Alphabet: 234567abcdefghijklmnopqrstuvwxyz';
diff --git a/migrations/2025-11-03-223026_add_post_status_enum/down.sql b/migrations/2025-11-03-223026_add_post_status_enum/down.sql
deleted file mode 100644
index 58012d60..00000000
--- a/migrations/2025-11-03-223026_add_post_status_enum/down.sql
+++ /dev/null
@@ -1,10 +0,0 @@
--- Revert post_status enum addition
-
--- Drop the index
-DROP INDEX IF EXISTS idx_posts_status;
-
--- Drop the status column
-ALTER TABLE posts DROP COLUMN IF EXISTS status;
-
--- Drop the enum type
-DROP TYPE IF EXISTS post_status;
diff --git a/migrations/2025-11-03-223026_add_post_status_enum/up.sql b/migrations/2025-11-03-223026_add_post_status_enum/up.sql
deleted file mode 100644
index 6ce0d751..00000000
--- a/migrations/2025-11-03-223026_add_post_status_enum/up.sql
+++ /dev/null
@@ -1,20 +0,0 @@
--- Add post_status enum type
--- This replaces the unresolved_refs table with a status column on posts
-
-CREATE TYPE post_status AS ENUM (
-    'complete',      -- Normal post with full content
-    'stub',          -- Placeholder for referenced post that doesn't exist yet
-    'deleted',       -- Soft-deleted post
-    'forbidden'      -- Post exists but is forbidden to access
-);
-
--- Add status column to posts table
-ALTER TABLE posts
-ADD COLUMN status post_status NOT NULL DEFAULT 'complete';
-
--- Create index on status column (partial index for non-complete posts)
-CREATE INDEX idx_posts_status ON posts(status) WHERE status != 'complete';
-
--- Add comment for documentation
-COMMENT ON COLUMN posts.status IS
-'Post lifecycle state: complete (normal), stub (placeholder), deleted (soft-deleted), forbidden (access denied)';
diff --git a/migrations/2025-11-04-060024_add_maintain_postgates_function/down.sql b/migrations/2025-11-04-060024_add_maintain_postgates_function/down.sql
deleted file mode 100644
index d9e94e24..00000000
--- a/migrations/2025-11-04-060024_add_maintain_postgates_function/down.sql
+++ /dev/null
@@ -1,2 +0,0 @@
--- Remove maintain_postgates function
-DROP FUNCTION IF EXISTS maintain_postgates(TEXT, TEXT[], TIMESTAMP);
diff --git a/migrations/2025-11-04-060024_add_maintain_postgates_function/up.sql b/migrations/2025-11-04-060024_add_maintain_postgates_function/up.sql
deleted file mode 100644
index 214821b4..00000000
--- a/migrations/2025-11-04-060024_add_maintain_postgates_function/up.sql
+++ /dev/null
@@ -1,70 +0,0 @@
--- Create maintain_postgates function
--- Updates the 'detached' flag on existing post_embed_record entries when a postgate changes
---
--- Parameters:
--- $1: post_uri - AT URI of the post that has the postgate (e.g., "at://did:plc:abc/app.bsky.feed.post/xyz")
--- $2: detached_uris - Array of post URIs that should be marked as detached
--- $3: disable_effective - Timestamp when the global disable rule became effective (nullable)
---
--- Returns: Number of rows updated (as BIGINT for compatibility with conn.execute)
-CREATE OR REPLACE FUNCTION maintain_postgates(
-    post_uri TEXT,
-    detached_uris TEXT[],
-    disable_effective TIMESTAMP
-) RETURNS BIGINT AS $$
-DECLARE
-    target_post_id BIGINT;
-    rows_updated BIGINT;
-BEGIN
-    -- Look up the post_id from the post URI
-    SELECT p.id INTO target_post_id
-    FROM posts p
-    INNER JOIN actors a ON p.actor_id = a.id
-    WHERE 'at://' || a.did || '/app.bsky.feed.post/' || i64_to_tid(p.rkey) = post_uri;
-
-    -- If post doesn't exist, nothing to do
-    IF target_post_id IS NULL THEN
-        RETURN 0;
-    END IF;
-
-    -- Update post_embed_record entries to mark quotes as detached
-    -- A quote post is detached if:
-    -- 1. It's in the detached_uris list, OR
-    -- 2. The global disable rule exists (disable_effective is not null) AND the quoting post was created after disable_effective
-    WITH detached_post_ids AS (
-        -- Resolve the detached URIs to post IDs
-        SELECT p.id
-        FROM UNNEST(detached_uris) AS uri
-        CROSS JOIN LATERAL (
-            SELECT
-                SPLIT_PART(SUBSTRING(uri FROM 6), '/', 1) as did,
-                SPLIT_PART(SUBSTRING(uri FROM 6), '/', 3) as rkey
-        ) parts
-        INNER JOIN actors a ON a.did = parts.did
-        INNER JOIN posts p ON p.actor_id = a.id AND p.rkey = tid_to_i64(parts.rkey)
-    ),
-    posts_to_update AS (
-        -- Find all posts that quote the target post
-        SELECT per.post_id
-        FROM post_embed_record per
-        WHERE per.embedded_post_id = target_post_id
-    )
-    UPDATE post_embed_record per
-    SET detached = (
-        -- Mark as detached if either:
-        -- 1. The quoting post is in the specific detached list
-        per.post_id IN (SELECT id FROM detached_post_ids)
-        OR
-        -- 2. Global disable rule applies (quote created after disable_effective)
-        (disable_effective IS NOT NULL AND (
-            SELECT tid_timestamp(p.rkey)
-            FROM posts p
-            WHERE p.id = per.post_id
-        ) > disable_effective)
-    )
-    WHERE per.embedded_post_id = target_post_id;
-
-    GET DIAGNOSTICS rows_updated = ROW_COUNT;
-    RETURN rows_updated;
-END;
-$$ LANGUAGE plpgsql;
diff --git a/migrations/2025-11-04-063554_make_post_content_nullable/down.sql b/migrations/2025-11-04-063554_make_post_content_nullable/down.sql
deleted file mode 100644
index 1a45ee7b..00000000
--- a/migrations/2025-11-04-063554_make_post_content_nullable/down.sql
+++ /dev/null
@@ -1,9 +0,0 @@
--- Revert content to NOT NULL
--- Set empty string for any NULL values before adding constraint
-
-UPDATE posts
-SET content = ''
-WHERE content IS NULL;
-
-ALTER TABLE posts
-ALTER COLUMN content SET NOT NULL;
diff --git a/migrations/2025-11-04-063554_make_post_content_nullable/up.sql b/migrations/2025-11-04-063554_make_post_content_nullable/up.sql
deleted file mode 100644
index e4b60ca6..00000000
--- a/migrations/2025-11-04-063554_make_post_content_nullable/up.sql
+++ /dev/null
@@ -1,9 +0,0 @@
--- Make post content nullable
--- This allows stubs (status='stub') and deleted posts (status='deleted') to have NULL content
--- instead of empty strings, which is more semantically correct.
-
-ALTER TABLE posts
-ALTER COLUMN content DROP NOT NULL;
-
-COMMENT ON COLUMN posts.content IS
-'Post text content. NULL for stub posts (not yet fetched) or deleted posts (soft-deleted).';
diff --git a/migrations/2025-11-04-064359_add_feedgen_labeler_status/down.sql b/migrations/2025-11-04-064359_add_feedgen_labeler_status/down.sql
deleted file mode 100644
index 1643b8ed..00000000
--- a/migrations/2025-11-04-064359_add_feedgen_labeler_status/down.sql
+++ /dev/null
@@ -1,18 +0,0 @@
--- Revert feedgen.name to NOT NULL (set empty string for NULL values)
-UPDATE feedgens
-SET name = ''
-WHERE name IS NULL;
-
-ALTER TABLE feedgens
-ALTER COLUMN name SET NOT NULL;
-
--- Drop status columns
-ALTER TABLE labelers
-DROP COLUMN status;
-
-ALTER TABLE feedgens
-DROP COLUMN status;
-
--- Drop status enums
-DROP TYPE labeler_status;
-DROP TYPE feedgen_status;
diff --git a/migrations/2025-11-04-064359_add_feedgen_labeler_status/up.sql b/migrations/2025-11-04-064359_add_feedgen_labeler_status/up.sql
deleted file mode 100644
index 3151f656..00000000
--- a/migrations/2025-11-04-064359_add_feedgen_labeler_status/up.sql
+++ /dev/null
@@ -1,24 +0,0 @@
--- Create status enums for feedgens and labelers
-CREATE TYPE feedgen_status AS ENUM ('complete', 'stub', 'deleted');
-CREATE TYPE labeler_status AS ENUM ('complete', 'stub', 'deleted');
-
--- Add status columns with default 'complete' for existing records
-ALTER TABLE feedgens
-ADD COLUMN status feedgen_status NOT NULL DEFAULT 'complete';
-
-ALTER TABLE labelers
-ADD COLUMN status labeler_status NOT NULL DEFAULT 'complete';
-
--- Make feedgen.name nullable (NULL for stubs)
-ALTER TABLE feedgens
-ALTER COLUMN name DROP NOT NULL;
-
--- Add comments explaining stub usage
-COMMENT ON COLUMN feedgens.status IS
-'Status of the feedgen record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = soft-deleted.';
-
-COMMENT ON COLUMN labelers.status IS
-'Status of the labeler record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = soft-deleted.';
-
-COMMENT ON COLUMN feedgens.name IS
-'Feedgen display name. NULL for stub feedgens (not yet fetched) or deleted feedgens (soft-deleted).';
diff --git a/migrations/2025-11-04-133658_remove_unresolved_refs/down.sql b/migrations/2025-11-04-133658_remove_unresolved_refs/down.sql
deleted file mode 100644
index e34d4eae..00000000
--- a/migrations/2025-11-04-133658_remove_unresolved_refs/down.sql
+++ /dev/null
@@ -1,26 +0,0 @@
--- Recreate unresolved_reason enum
-CREATE TYPE unresolved_reason AS ENUM ('pending', 'deleted', 'forbidden', 'not_found');
-
--- Recreate unresolved_refs table
-CREATE TABLE unresolved_refs (
-    id SERIAL PRIMARY KEY,
-    cid BYTEA NOT NULL UNIQUE,
-    reason unresolved_reason NOT NULL,
-    reference_count INTEGER NOT NULL DEFAULT 1,
-    actor_did TEXT NOT NULL,
-    rkey BIGINT NOT NULL,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    last_checked TIMESTAMPTZ
-);
-
-CREATE INDEX idx_unresolved_refs_cid ON unresolved_refs(cid);
-CREATE INDEX idx_unresolved_refs_reason ON unresolved_refs(reason);
-CREATE INDEX idx_unresolved_refs_last_checked ON unresolved_refs(last_checked) WHERE last_checked IS NOT NULL;
-
--- Add back unresolved reference columns to posts table
-ALTER TABLE posts
-ADD COLUMN unresolved_parent_id INTEGER REFERENCES unresolved_refs(id),
-ADD COLUMN unresolved_root_id INTEGER REFERENCES unresolved_refs(id);
-
-CREATE INDEX idx_posts_unresolved_parent ON posts(unresolved_parent_id) WHERE unresolved_parent_id IS NOT NULL;
-CREATE INDEX idx_posts_unresolved_root ON posts(unresolved_root_id) WHERE unresolved_root_id IS NOT NULL;
diff --git a/migrations/2025-11-04-133658_remove_unresolved_refs/up.sql b/migrations/2025-11-04-133658_remove_unresolved_refs/up.sql
deleted file mode 100644
index 91bd1490..00000000
--- a/migrations/2025-11-04-133658_remove_unresolved_refs/up.sql
+++ /dev/null
@@ -1,11 +0,0 @@
--- Remove unresolved reference columns from posts table
--- These columns are replaced by the stub pattern
-ALTER TABLE posts
-DROP COLUMN unresolved_parent_id,
-DROP COLUMN unresolved_root_id;
-
--- Drop the unresolved_refs table (unused - 0 rows)
-DROP TABLE IF EXISTS unresolved_refs;
-
--- Drop the unresolved_reason enum type
-DROP TYPE IF EXISTS unresolved_reason;
diff --git a/migrations/2025-11-04-154903_remove_pending_cid_columns/down.sql b/migrations/2025-11-04-154903_remove_pending_cid_columns/down.sql
deleted file mode 100644
index 963028e1..00000000
--- a/migrations/2025-11-04-154903_remove_pending_cid_columns/down.sql
+++ /dev/null
@@ -1,27 +0,0 @@
--- Recreate pending CID columns (for rollback purposes)
-
--- Add columns to likes table
-ALTER TABLE likes
-ADD COLUMN pending_subject_cid BYTEA,
-ADD COLUMN pending_via_post_cid BYTEA;
-
--- Add columns to reposts table
-ALTER TABLE reposts
-ADD COLUMN pending_post_cid BYTEA,
-ADD COLUMN pending_via_post_cid BYTEA;
-
--- Add columns to threadgates table
-ALTER TABLE threadgates
-ADD COLUMN pending_post_cid BYTEA;
-
--- Add columns to postgates table
-ALTER TABLE postgates
-ADD COLUMN pending_post_cid BYTEA;
-
--- Recreate indexes
-CREATE INDEX idx_likes_pending_subject_cid ON likes(pending_subject_cid) WHERE pending_subject_cid IS NOT NULL;
-CREATE INDEX idx_likes_pending_via_post_cid ON likes(pending_via_post_cid) WHERE pending_via_post_cid IS NOT NULL;
-CREATE INDEX idx_reposts_pending_post_cid ON reposts(pending_post_cid) WHERE pending_post_cid IS NOT NULL;
-CREATE INDEX idx_reposts_pending_via_post_cid ON reposts(pending_via_post_cid) WHERE pending_via_post_cid IS NOT NULL;
-CREATE INDEX idx_threadgates_pending_post_cid ON threadgates(pending_post_cid) WHERE pending_post_cid IS NOT NULL;
-CREATE INDEX idx_postgates_pending_post_cid ON postgates(pending_post_cid) WHERE pending_post_cid IS NOT NULL;
diff --git a/migrations/2025-11-04-154903_remove_pending_cid_columns/up.sql b/migrations/2025-11-04-154903_remove_pending_cid_columns/up.sql
deleted file mode 100644
index 7ca5342f..00000000
--- a/migrations/2025-11-04-154903_remove_pending_cid_columns/up.sql
+++ /dev/null
@@ -1,28 +0,0 @@
--- Remove pending CID columns from likes, reposts, threadgates, and postgates tables
--- These columns are replaced by the stub pattern where we create stub records immediately
-
--- Drop indexes first
-DROP INDEX IF EXISTS idx_likes_pending_subject_cid;
-DROP INDEX IF EXISTS idx_likes_pending_via_post_cid;
-DROP INDEX IF EXISTS idx_reposts_pending_post_cid;
-DROP INDEX IF EXISTS idx_reposts_pending_via_post_cid;
-DROP INDEX IF EXISTS idx_threadgates_pending_post_cid;
-DROP INDEX IF EXISTS idx_postgates_pending_post_cid;
-
--- Drop columns from likes table
-ALTER TABLE likes
-DROP COLUMN IF EXISTS pending_subject_cid,
-DROP COLUMN IF EXISTS pending_via_post_cid;
-
--- Drop columns from reposts table
-ALTER TABLE reposts
-DROP COLUMN IF EXISTS pending_post_cid,
-DROP COLUMN IF EXISTS pending_via_post_cid;
-
--- Drop columns from threadgates table
-ALTER TABLE threadgates
-DROP COLUMN IF EXISTS pending_post_cid;
-
--- Drop columns from postgates table
-ALTER TABLE postgates
-DROP COLUMN IF EXISTS pending_post_cid;
diff --git a/migrations/2025-11-04-190708_drop_notification_tables/down.sql b/migrations/2025-11-04-190708_drop_notification_tables/down.sql
deleted file mode 100644
index b1dcb9cd..00000000
--- a/migrations/2025-11-04-190708_drop_notification_tables/down.sql
+++ /dev/null
@@ -1,22 +0,0 @@
--- Recreate notification tables and types (for rollback)
-
-CREATE TYPE notification_reason AS ENUM ('follow', 'like', 'repost', 'reply', 'mention', 'quote');
-CREATE TYPE notification_record_type AS ENUM ('post', 'like', 'repost', 'follow', 'block');
-
-CREATE TABLE notification_seens (
-    actor_id INT4 PRIMARY KEY REFERENCES actors(id),
-    seen_at TIMESTAMPTZ NOT NULL
-);
-
-CREATE TABLE notifications (
-    recipient_actor_id INT4 NOT NULL REFERENCES actors(id),
-    record_type notification_record_type NOT NULL,
-    record_id INT8 NOT NULL,
-    reason notification_reason NOT NULL,
-    reason_subject_post_id INT8 REFERENCES posts(id),
-    is_read BOOLEAN NOT NULL DEFAULT FALSE,
-    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-    PRIMARY KEY (recipient_actor_id, record_type, record_id)
-);
-
-CREATE INDEX notifications_recipient_indexed_at_idx ON notifications (recipient_actor_id, created_at DESC);
diff --git a/migrations/2025-11-04-190708_drop_notification_tables/up.sql b/migrations/2025-11-04-190708_drop_notification_tables/up.sql
deleted file mode 100644
index 091572d0..00000000
--- a/migrations/2025-11-04-190708_drop_notification_tables/up.sql
+++ /dev/null
@@ -1,12 +0,0 @@
--- Drop notification tables and types
---
--- NOTE: This migration removes notification tables from the database schema.
--- Notifications are now stored in Redis for better performance and automatic TTL-based cleanup.
---
--- Consumer writes to Redis via redis_notifications module.
--- Parakeet AppView will read from Redis (to be implemented in separate PR).
-
-DROP TABLE IF EXISTS notifications;
-DROP TABLE IF EXISTS notification_seens;
-DROP TYPE IF EXISTS notification_record_type;
-DROP TYPE IF EXISTS notification_reason;
diff --git a/migrations/2025-11-05-021106_add_list_starterpack_status/down.sql b/migrations/2025-11-05-021106_add_list_starterpack_status/down.sql
deleted file mode 100644
index 67de7c2e..00000000
--- a/migrations/2025-11-05-021106_add_list_starterpack_status/down.sql
+++ /dev/null
@@ -1,32 +0,0 @@
--- Restore list_items.list_id to nullable
-ALTER TABLE list_items
-ALTER COLUMN list_id DROP NOT NULL;
-
--- Restore list_items.pending_list_cid column
-ALTER TABLE list_items
-ADD COLUMN pending_list_cid bytea;
-
--- Create index for pending list CID lookups
-CREATE INDEX idx_list_items_pending_list_cid
-ON list_items (pending_list_cid)
-WHERE pending_list_cid IS NOT NULL;
-
--- Restore NOT NULL constraints
-ALTER TABLE starterpacks
-ALTER COLUMN name SET NOT NULL;
-
-ALTER TABLE lists
-ALTER COLUMN list_type SET NOT NULL,
-ALTER COLUMN name SET NOT NULL;
-
--- Remove status columns
-ALTER TABLE starterpacks
-DROP COLUMN status;
-
-ALTER TABLE lists
-DROP COLUMN status;
-
--- Drop status enum types (from the original migration that was run)
-DROP TYPE IF EXISTS record_status;
-DROP TYPE IF EXISTS starterpack_status;
-DROP TYPE IF EXISTS list_status;
diff --git a/migrations/2025-11-05-021106_add_list_starterpack_status/up.sql b/migrations/2025-11-05-021106_add_list_starterpack_status/up.sql
deleted file mode 100644
index 6f5a366a..00000000
--- a/migrations/2025-11-05-021106_add_list_starterpack_status/up.sql
+++ /dev/null
@@ -1,53 +0,0 @@
--- Create a shared record_status enum type for all record tables
--- This replaces the pattern of creating separate enums (post_status, feedgen_status, etc.)
--- All tables can share this single enum type
-CREATE TYPE record_status AS ENUM ('complete', 'stub', 'deleted', 'forbidden');
-
--- Add status columns with default 'complete' for existing records
-ALTER TABLE lists
-ADD COLUMN status record_status NOT NULL DEFAULT 'complete';
-
-ALTER TABLE starterpacks
-ADD COLUMN status record_status NOT NULL DEFAULT 'complete';
-
--- Make list fields nullable for stub support
--- Stub lists don't have list_type or name until fetched
-ALTER TABLE lists
-ALTER COLUMN list_type DROP NOT NULL,
-ALTER COLUMN name DROP NOT NULL;
-
--- Make starterpack fields nullable for stub support
--- Stub starterpacks don't have name until fetched
-ALTER TABLE starterpacks
-ALTER COLUMN name DROP NOT NULL;
-
--- Delete orphaned list_items with NULL list_id
--- These are list items referencing lists that don't exist yet
--- With the new stubbing system, these will be recreated with stub lists
-DELETE FROM list_items WHERE list_id IS NULL;
-
--- Remove list_items.pending_list_cid column (replaced by list stubbing)
--- With list stubbing, we can now create stub lists and reference them directly
--- instead of storing a pending CID
-ALTER TABLE list_items
-DROP COLUMN IF EXISTS pending_list_cid;
-
--- Make list_items.list_id NOT NULL since we can now always reference a list (stub or complete)
-ALTER TABLE list_items
-ALTER COLUMN list_id SET NOT NULL;
-
--- Add comments explaining status usage
-COMMENT ON COLUMN lists.status IS
-'Status of the list record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = user deleted, ''forbidden'' = removed by moderators.';
-
-COMMENT ON COLUMN lists.list_type IS
-'Type of list (e.g., modlist, curatelist). NULL for stub lists (not yet fetched).';
-
-COMMENT ON COLUMN lists.name IS
-'List display name. NULL for stub lists (not yet fetched).';
-
-COMMENT ON COLUMN starterpacks.status IS
-'Status of the starter pack record. ''complete'' = full data, ''stub'' = placeholder (not yet fetched), ''deleted'' = user deleted, ''forbidden'' = removed by moderators.';
-
-COMMENT ON COLUMN starterpacks.name IS
-'Starter pack name. NULL for stub starter packs (not yet fetched).';
diff --git a/migrations/2025-11-05-055339_remove_reposts_actor_post_unique_constraint/down.sql b/migrations/2025-11-05-055339_remove_reposts_actor_post_unique_constraint/down.sql
deleted file mode 100644
index 7498bb1c..00000000
--- a/migrations/2025-11-05-055339_remove_reposts_actor_post_unique_constraint/down.sql
+++ /dev/null
@@ -1,4 +0,0 @@
--- Restore the UNIQUE (actor_id, post_id) constraint.
--- Note: This may fail if there are now multiple reposts of the same post by the same actor.
-
-ALTER TABLE reposts ADD CONSTRAINT reposts_actor_id_post_id_key UNIQUE (actor_id, post_id);
diff --git a/migrations/2025-11-05-055339_remove_reposts_actor_post_unique_constraint/up.sql b/migrations/2025-11-05-055339_remove_reposts_actor_post_unique_constraint/up.sql
deleted file mode 100644
index 474e6e12..00000000
--- a/migrations/2025-11-05-055339_remove_reposts_actor_post_unique_constraint/up.sql
+++ /dev/null
@@ -1,8 +0,0 @@
--- Remove the incorrect UNIQUE (actor_id, post_id) constraint from reposts table.
---
--- In the AT Protocol, users CAN create multiple reposts of the same post with different rkeys.
--- Each repost is identified by its AT-URI: at://did/app.bsky.feed.repost/{rkey}
---
--- The only uniqueness constraint should be (actor_id, rkey), which identifies the repost record itself.
-
-ALTER TABLE reposts DROP CONSTRAINT reposts_actor_id_post_id_key;
diff --git a/migrations/2025-11-05-120000_add_account_created_at_to_actors/down.sql b/migrations/2025-11-05-120000_add_account_created_at_to_actors/down.sql
deleted file mode 100644
index 2b92ccf7..00000000
--- a/migrations/2025-11-05-120000_add_account_created_at_to_actors/down.sql
+++ /dev/null
@@ -1,3 +0,0 @@
--- Remove account_created_at column from actors table
-DROP INDEX IF EXISTS idx_actors_account_created_at;
-ALTER TABLE actors DROP COLUMN account_created_at;
diff --git a/migrations/2025-11-05-120000_add_account_created_at_to_actors/up.sql b/migrations/2025-11-05-120000_add_account_created_at_to_actors/up.sql
deleted file mode 100644
index 924995e2..00000000
--- a/migrations/2025-11-05-120000_add_account_created_at_to_actors/up.sql
+++ /dev/null
@@ -1,8 +0,0 @@
--- Add account_created_at column to actors table
--- This represents when the account was created (from PLC directory first operation)
--- Nullable because we may not know it for all actors initially (especially stubs)
--- Will be populated during handle resolution
-ALTER TABLE actors ADD COLUMN account_created_at TIMESTAMP WITH TIME ZONE;
-
--- Create index for potential queries by account age
-CREATE INDEX idx_actors_account_created_at ON actors(account_created_at) WHERE account_created_at IS NOT NULL;
diff --git a/migrations/2025-11-07-192945_posts_content_compression/down.sql b/migrations/2025-11-07-192945_posts_content_compression/down.sql
deleted file mode 100644
index 8063b4fc..00000000
--- a/migrations/2025-11-07-192945_posts_content_compression/down.sql
+++ /dev/null
@@ -1,19 +0,0 @@
--- Revert post content compression
--- Changes posts.content back from BYTEA to TEXT
--- Removes content_version and search_vector columns
-
--- Drop index first
-DROP INDEX IF EXISTS idx_posts_search_vector;
-
--- Drop CHECK constraint
-ALTER TABLE posts
-    DROP CONSTRAINT IF EXISTS posts_content_version_check;
-
--- Drop columns if they exist (use IF EXISTS for idempotency)
-ALTER TABLE posts
-    DROP COLUMN IF EXISTS content_version,
-    DROP COLUMN IF EXISTS search_vector;
-
--- Revert content column from BYTEA to TEXT
-ALTER TABLE posts
-    ALTER COLUMN content TYPE TEXT;
diff --git a/migrations/2025-11-07-192945_posts_content_compression/up.sql b/migrations/2025-11-07-192945_posts_content_compression/up.sql
deleted file mode 100644
index 51a36ea3..00000000
--- a/migrations/2025-11-07-192945_posts_content_compression/up.sql
+++ /dev/null
@@ -1,26 +0,0 @@
--- Add post content compression
--- Changes posts.content from TEXT to BYTEA (zstd-compressed)
--- Adds content_version to track dictionary version
--- Adds search_vector for selective full-text search on allowlisted users
-
--- Alter posts table: TEXT -> BYTEA
-ALTER TABLE posts
-    ALTER COLUMN content TYPE BYTEA USING NULL,
-    ADD COLUMN content_version SMALLINT,
-    ADD COLUMN search_vector tsvector;
-
-COMMENT ON COLUMN posts.content IS 'Zstd-compressed post content (BYTEA)';
-COMMENT ON COLUMN posts.content_version IS 'Dictionary version (NULL if no content)';
-COMMENT ON COLUMN posts.search_vector IS 'Full-text search vector (NULL except for allowlisted users)';
-
--- Add CHECK constraint: both NULL or both set
-ALTER TABLE posts
-    ADD CONSTRAINT posts_content_version_check
-    CHECK (
-        (content IS NULL AND content_version IS NULL) OR
-        (content IS NOT NULL AND content_version IS NOT NULL)
-    );
-
--- Create GIN index for search_vector (sparse index since most are NULL)
-CREATE INDEX idx_posts_search_vector ON posts USING GIN (search_vector)
-    WHERE search_vector IS NOT NULL;
diff --git a/migrations/2025-11-08-001834_drop_tid_created_at_columns/down.sql b/migrations/2025-11-08-001834_drop_tid_created_at_columns/down.sql
deleted file mode 100644
index d49c2a8d..00000000
--- a/migrations/2025-11-08-001834_drop_tid_created_at_columns/down.sql
+++ /dev/null
@@ -1,49 +0,0 @@
--- Rollback: Re-add created_at columns and populate from TID rkeys
-
--- Add created_at columns back (nullable initially)
-ALTER TABLE posts ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE postgates ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE likes ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE reposts ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE follows ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE blocks ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE bookmarks ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE lists ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE list_items ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE list_blocks ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE starterpacks ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE threadgates ADD COLUMN created_at TIMESTAMPTZ;
-ALTER TABLE verification ADD COLUMN created_at TIMESTAMPTZ;
-
--- Populate from TID rkeys using helper function
-UPDATE posts SET created_at = tid_timestamp(rkey);
-UPDATE postgates SET created_at = tid_timestamp(rkey);
-UPDATE likes SET created_at = tid_timestamp(rkey);
-UPDATE reposts SET created_at = tid_timestamp(rkey);
-UPDATE follows SET created_at = tid_timestamp(rkey);
-UPDATE blocks SET created_at = tid_timestamp(rkey);
-UPDATE bookmarks SET created_at = tid_timestamp(rkey);
-UPDATE lists SET created_at = tid_timestamp(rkey);
-UPDATE list_items SET created_at = tid_timestamp(rkey);
-UPDATE list_blocks SET created_at = tid_timestamp(rkey);
-UPDATE starterpacks SET created_at = tid_timestamp(rkey);
-UPDATE threadgates SET created_at = tid_timestamp(rkey);
-UPDATE verification SET created_at = tid_timestamp(rkey);
-
--- Set NOT NULL constraints
-ALTER TABLE posts ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE postgates ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE likes ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE reposts ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE follows ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE blocks ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE bookmarks ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE lists ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE list_items ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE list_blocks ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE starterpacks ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE threadgates ALTER COLUMN created_at SET NOT NULL;
-ALTER TABLE verification ALTER COLUMN created_at SET NOT NULL;
-
--- Drop the helper function
-DROP FUNCTION IF EXISTS tid_timestamp(BIGINT);
diff --git a/migrations/2025-11-08-001834_drop_tid_created_at_columns/up.sql b/migrations/2025-11-08-001834_drop_tid_created_at_columns/up.sql
deleted file mode 100644
index 9aefc57e..00000000
--- a/migrations/2025-11-08-001834_drop_tid_created_at_columns/up.sql
+++ /dev/null
@@ -1,30 +0,0 @@
--- Drop created_at column from TID-based tables
--- TIDs encode microsecond timestamps in their upper 53 bits, making created_at redundant
--- Storage savings: ~140 GB at network scale (2B posts, 11.5B likes, 3B follows, 600M reposts)
-
--- Add PostgreSQL helper function to extract timestamp from TID
-CREATE OR REPLACE FUNCTION tid_timestamp(tid_int8 BIGINT)
-RETURNS TIMESTAMPTZ AS $$
-BEGIN
-    -- Extract timestamp (53 bits) by right shifting 10 bits (clock ID)
-    -- Convert microseconds to PostgreSQL timestamp
-    RETURN to_timestamp((tid_int8 >> 10) / 1000000.0);
-END;
-$$ LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
-
--- Drop created_at columns from tables with TID rkeys
--- These timestamps can be derived from the rkey using tid_timestamp(rkey)
-
-ALTER TABLE posts DROP COLUMN created_at;
-ALTER TABLE postgates DROP COLUMN created_at;
-ALTER TABLE likes DROP COLUMN created_at;
-ALTER TABLE reposts DROP COLUMN created_at;
-ALTER TABLE follows DROP COLUMN created_at;
-ALTER TABLE blocks DROP COLUMN created_at;
-ALTER TABLE bookmarks DROP COLUMN created_at;
-ALTER TABLE lists DROP COLUMN created_at;
-ALTER TABLE list_items DROP COLUMN created_at;
-ALTER TABLE list_blocks DROP COLUMN created_at;
-ALTER TABLE starterpacks DROP COLUMN created_at;
-ALTER TABLE threadgates DROP COLUMN created_at;
-ALTER TABLE verification DROP COLUMN created_at;
diff --git a/parakeet-db/src/schema.rs b/parakeet-db/src/schema.rs
index ea30c65e..02ee6ec1 100644
--- a/parakeet-db/src/schema.rs
+++ b/parakeet-db/src/schema.rs
@@ -447,6 +447,7 @@ diesel::table! {
         rkey -> Int8,
         cid -> Bytea,
         content -> Nullable<Bytea>,
+        content_version -> Nullable<Int2>,
         language_code -> Nullable<LanguageCode>,
         tags -> Array<Nullable<Text>>,
         parent_post_id -> Nullable<Int8>,
@@ -455,7 +456,6 @@ diesel::table! {
         embed_subtype -> Nullable<EmbedType>,
         violates_threadgate -> Bool,
         status -> PostStatus,
-        content_version -> Nullable<Int2>,
         search_vector -> Nullable<Tsvector>,
     }
 }