diff --git a/apps/obsidian/src/utils/conceptConversion.ts b/apps/obsidian/src/utils/conceptConversion.ts
index 339311234..931ff36ab 100644
--- a/apps/obsidian/src/utils/conceptConversion.ts
+++ b/apps/obsidian/src/utils/conceptConversion.ts
@@ -10,6 +10,7 @@ import type { DiscourseNodeInVault } from "./getDiscourseNodes";
 import type { LocalConceptDataInput } from "@repo/database/inputTypes";
 import type { ObsidianDiscourseNodeData } from "./syncDgNodesToSupabase";
 import type { Json } from "@repo/database/dbTypes";
+import { extractContentFromTitle } from "./extractContentFromTitle";
 
 /**
  * Get extra data (author, timestamps) from file metadata
@@ -160,12 +161,17 @@ export const discourseRelationTripleSchemaToLocalConcept = ({
 export const discourseNodeInstanceToLocalConcept = (
   context: SupabaseContext,
   nodeData: ObsidianDiscourseNodeData,
+  nodeTypesById: Record<string, DiscourseNode>,
 ): LocalConceptDataInput => {
   const extraData = getNodeExtraData(nodeData.file, context.userId);
   const { nodeInstanceId, nodeTypeId, importedFromRid, ...otherData } =
     nodeData.frontmatter;
   const literal_content: Record<string, Json> = {
     label: nodeData.file.basename,
+    core_title: extractContentFromTitle(
+      nodeTypesById[nodeData.nodeTypeId]?.format ?? "",
+      nodeData.file.basename,
+    ),
     source_data: otherData as unknown as Json,
   };
   if (importedFromRid && typeof importedFromRid === "string")
diff --git a/apps/obsidian/src/utils/syncDgNodesToSupabase.ts b/apps/obsidian/src/utils/syncDgNodesToSupabase.ts
index 9eb73e40a..25207a24d 100644
--- a/apps/obsidian/src/utils/syncDgNodesToSupabase.ts
+++ b/apps/obsidian/src/utils/syncDgNodesToSupabase.ts
@@ -30,6 +30,10 @@ import { isAcceptedSchema } from "./typeUtils";
 import { getTemplatePluginInfo } from "./templates";
 import { difference } from "@repo/utils/setOperations";
 import { getAllPages } from "@repo/database/lib/pagination";
+import {
+  CORE_TITLE_PROBE_SELECT,
+  partitionByCoreTitle,
+} from "@repo/database/lib/coreTitleBackfill";
 
 const DEFAULT_TIME = "1970-01-01";
 export type ChangeType = "title" | "content";
@@ -231,6 +235,17 @@ type BuildChangedNodesOptions = {
   fullSync?: boolean;
 };
 
+type CoreTitleBackfillCounts = {
+  backfilled: number;
+  skipped: number;
+  orphaned: number;
+};
+
+type BuildChangedNodesResult = {
+  changedNodes: ObsidianDiscourseNodeData[];
+  coreTitleBackfill: CoreTitleBackfillCounts | null;
+};
+
 const mergeChangeTypes = (
   base: ChangeType[],
   additional: ChangeType[],
@@ -320,15 +335,33 @@ const detectNodeChanges = (
   return changeTypes;
 };
 
+const noticeCoreTitleBackfill = ({
+  backfilled,
+  skipped,
+  orphaned,
+}: CoreTitleBackfillCounts): void => {
+  if (backfilled === 0 && orphaned === 0) return;
+  const messages = [
+    `Backfilled core title for ${backfilled} node${backfilled === 1 ? "" : "s"}.`,
+    `${skipped} already had one.`,
+  ];
+  if (orphaned > 0) {
+    messages.push(
+      `${orphaned} no longer match a discourse node in this vault.`,
+    );
+  }
+  new Notice(messages.join(" "), 5000);
+};
+
 const buildChangedNodesFromNodes = async ({
   nodes,
   supabaseClient,
   context,
   changeTypesByPath,
   fullSync = false,
-}: BuildChangedNodesOptions): Promise<ObsidianDiscourseNodeData[]> => {
+}: BuildChangedNodesOptions): Promise<BuildChangedNodesResult> => {
   if (nodes.length === 0) {
-    return [];
+    return { changedNodes: [], coreTitleBackfill: null };
   }
 
   const nodeInstanceIds = nodes.map((node) => node.nodeInstanceId);
@@ -344,11 +377,14 @@ const buildChangedNodesFromNodes = async ({
   );
   const changedNodes: ObsidianDiscourseNodeData[] = [];
   let missingConcepts: Set<string> | undefined;
+  let coreTitleProbe:
+    | { missingCoreTitleIds: Set<string>; skipped: number; orphaned: number }
+    | undefined;
   if (fullSync) {
     const existingConceptIds = await getAllPages(
       supabaseClient
         .from("my_concepts")
-        .select("source_local_id")
+        .select(CORE_TITLE_PROBE_SELECT)
         .eq("space_id", context.spaceId)
         .eq("arity", 0)
         .eq("is_schema", false)
@@ -369,6 +405,13 @@ const buildChangedNodesFromNodes = async ({
           .filter((id) => id !== null),
       );
       missingConcepts = difference(nodeIds, dbConceptIds);
+      const { missingCoreTitleIds, withCoreTitleCount } =
+        partitionByCoreTitle(existingConceptIds);
+      coreTitleProbe = {
+        missingCoreTitleIds,
+        skipped: withCoreTitleCount,
+        orphaned: difference(missingCoreTitleIds, nodeIds).size,
+      };
     }
   }
 
@@ -389,7 +432,8 @@ const buildChangedNodesFromNodes = async ({
 
     if (
       finalChangeTypes.length === 0 &&
-      !missingConcepts?.has(node.nodeInstanceId)
+      !missingConcepts?.has(node.nodeInstanceId) &&
+      !coreTitleProbe?.missingCoreTitleIds.has(node.nodeInstanceId)
     ) {
       continue;
     }
@@ -405,7 +449,18 @@ const buildChangedNodesFromNodes = async ({
     });
   }
 
-  return changedNodes;
+  return {
+    changedNodes,
+    coreTitleBackfill: coreTitleProbe
+      ? {
+          backfilled: changedNodes.filter((node) =>
+            coreTitleProbe.missingCoreTitleIds.has(node.nodeInstanceId),
+          ).length,
+          skipped: coreTitleProbe.skipped,
+          orphaned: coreTitleProbe.orphaned,
+        }
+      : null,
+  };
 };
 
 export const syncAllNodesAndRelations = async (
@@ -426,14 +481,15 @@ export const syncAllNodesAndRelations = async (
 
     const allNodes = await collectDiscourseNodesFromVault(plugin, true);
 
-    const changedNodeInstances = relationsOnly
-      ? []
-      : await buildChangedNodesFromNodes({
-          nodes: allNodes,
-          supabaseClient,
-          context,
-          fullSync: true,
-        });
+    const { changedNodes: changedNodeInstances, coreTitleBackfill } =
+      relationsOnly
+        ? { changedNodes: [], coreTitleBackfill: null }
+        : await buildChangedNodesFromNodes({
+            nodes: allNodes,
+            supabaseClient,
+            context,
+            fullSync: true,
+          });
 
     const accountLocalId = plugin.settings.accountLocalId;
     if (!accountLocalId) {
@@ -458,8 +514,15 @@ export const syncAllNodesAndRelations = async (
       fullSync: true,
     });
 
+    if (coreTitleBackfill !== null) {
+      noticeCoreTitleBackfill(coreTitleBackfill);
+    }
+
     // When synced nodes are already published, ensure non-text assets are in storage.
-    await syncPublishedNodesAssets(plugin, changedNodeInstances);
+    await syncPublishedNodesAssets(
+      plugin,
+      changedNodeInstances.filter((node) => node.changeTypes.length > 0),
+    );
   } catch (error) {
     console.error("syncAllNodesAndRelations: Process failed:", error);
     throw error;
@@ -594,7 +657,7 @@ const convertDgToSupabaseConcepts = async ({
     .filter((n) => !!n);
 
   const nodeInstanceToLocalConcepts = nodesSince.map((node) => {
-    return discourseNodeInstanceToLocalConcept(context, node);
+    return discourseNodeInstanceToLocalConcept(context, node, nodeTypesById);
   });
 
   const relationInstancesData = await loadRelations(plugin);
@@ -935,7 +998,7 @@ export const syncDiscourseNodeChanges = async (
       return;
     }
 
-    const changedNodes = await buildChangedNodesFromNodes({
+    const { changedNodes } = await buildChangedNodesFromNodes({
       nodes: dgNodesInVault,
       supabaseClient,
       context,
diff --git a/apps/roam/src/utils/__tests__/coreTitleBackfill.test.ts b/apps/roam/src/utils/__tests__/coreTitleBackfill.test.ts
new file mode 100644
index 000000000..59eeddb45
--- /dev/null
+++ b/apps/roam/src/utils/__tests__/coreTitleBackfill.test.ts
@@ -0,0 +1,92 @@
+import { describe, expect, it } from "vitest";
+import {
+  buildCoreTitleBackfill,
+  mergeNodesBySourceLocalId,
+} from "../coreTitleBackfill";
+import type { RoamDiscourseNodeData } from "../getAllDiscourseNodesSince";
+
+const node = (sourceLocalId: string): RoamDiscourseNodeData => ({
+  author_local_id: "author",
+  author_name: "Author",
+  source_local_id: sourceLocalId,
+  created: "1",
+  last_modified: "2",
+  text: `CLM - ${sourceLocalId}`,
+  type: "claim-type",
+});
+
+describe("buildCoreTitleBackfill", () => {
+  it("forces in local nodes whose row has no core_title", () => {
+    const backfill = buildCoreTitleBackfill({
+      conceptRows: [
+        { source_local_id: "a", core_title: null },
+        { source_local_id: "b", core_title: "already set" },
+      ],
+      localNodes: [node("a"), node("b")],
+    });
+
+    expect(backfill.nodesToBackfill.map((n) => n.source_local_id)).toEqual([
+      "a",
+    ]);
+    expect(backfill.withCoreTitleCount).toBe(1);
+    expect(backfill.orphanedCount).toBe(0);
+  });
+
+  it("reports rows that are no longer in the graph as orphaned", () => {
+    const backfill = buildCoreTitleBackfill({
+      conceptRows: [
+        { source_local_id: "a", core_title: null },
+        { source_local_id: "gone", core_title: null },
+      ],
+      localNodes: [node("a")],
+    });
+
+    expect(backfill.nodesToBackfill.map((n) => n.source_local_id)).toEqual([
+      "a",
+    ]);
+    expect(backfill.orphanedCount).toBe(1);
+  });
+
+  it("skips rows without a source_local_id", () => {
+    const backfill = buildCoreTitleBackfill({
+      conceptRows: [{ source_local_id: null, core_title: null }],
+      localNodes: [node("a")],
+    });
+
+    expect(backfill.nodesToBackfill).toEqual([]);
+    expect(backfill.withCoreTitleCount).toBe(0);
+    expect(backfill.orphanedCount).toBe(0);
+  });
+
+  it("is a no-op once every row has a core_title", () => {
+    const backfill = buildCoreTitleBackfill({
+      conceptRows: [
+        { source_local_id: "a", core_title: "a" },
+        { source_local_id: "b", core_title: "b" },
+      ],
+      localNodes: [node("a"), node("b")],
+    });
+
+    expect(backfill.nodesToBackfill).toEqual([]);
+    expect(backfill.withCoreTitleCount).toBe(2);
+    expect(backfill.orphanedCount).toBe(0);
+  });
+});
+
+describe("mergeNodesBySourceLocalId", () => {
+  it("appends nodes that are not already in the batch", () => {
+    const merged = mergeNodesBySourceLocalId(
+      [node("a")],
+      [node("b"), node("c")],
+    );
+
+    expect(merged.map((n) => n.source_local_id)).toEqual(["a", "b", "c"]);
+  });
+
+  it("keeps the original node when both batches hold the same id", () => {
+    const original = node("a");
+    const merged = mergeNodesBySourceLocalId([original], [node("a")]);
+
+    expect(merged).toEqual([original]);
+  });
+});
diff --git a/apps/roam/src/utils/conceptConversion.ts b/apps/roam/src/utils/conceptConversion.ts
index 35c8f76a5..47d02803a 100644
--- a/apps/roam/src/utils/conceptConversion.ts
+++ b/apps/roam/src/utils/conceptConversion.ts
@@ -13,7 +13,7 @@ import getPageTitleByPageUid from "roamjs-components/queries/getPageTitleByPageU
 const getNodeExtraData = (
   node_uid: string,
 ): {
-  author_uid: string;
+  author_local_id: string;
   created: string;
   last_modified: string;
   page_uid: string;
@@ -50,7 +50,7 @@ const getNodeExtraData = (
   const created = new Date(created_t).toISOString();
   const last_modified = new Date(last_modified_t).toISOString();
   return {
-    author_uid,
+    author_local_id: author_uid,
     created,
     last_modified,
     page_uid,
@@ -202,7 +202,7 @@ export const discourseRelationDataToLocalConcept = (
   const created = new Date(
     Math.max(...nodeData.map((nd) => new Date(nd.created).getTime())),
   ).toISOString();
-  const author_local_id: string = nodeData[0].author_uid; // take any one; again until I get the relation object
+  const author_local_id: string = nodeData[0].author_local_id; // take any one; again until I get the relation object
   return {
     space_id: context.spaceId,
     source_local_id: relationUid,
diff --git a/apps/roam/src/utils/coreTitleBackfill.ts b/apps/roam/src/utils/coreTitleBackfill.ts
new file mode 100644
index 000000000..8794aee43
--- /dev/null
+++ b/apps/roam/src/utils/coreTitleBackfill.ts
@@ -0,0 +1,44 @@
+import { difference } from "@repo/utils/setOperations";
+import {
+  partitionByCoreTitle,
+  type CoreTitleProbeRow,
+} from "@repo/database/lib/coreTitleBackfill";
+import { type RoamDiscourseNodeData } from "./getAllDiscourseNodesSince";
+
+export type CoreTitleBackfill = {
+  nodesToBackfill: RoamDiscourseNodeData[];
+  withCoreTitleCount: number;
+  orphanedCount: number;
+};
+
+export const buildCoreTitleBackfill = ({
+  conceptRows,
+  localNodes,
+}: {
+  conceptRows: CoreTitleProbeRow[];
+  localNodes: RoamDiscourseNodeData[];
+}): CoreTitleBackfill => {
+  const { missingCoreTitleIds, withCoreTitleCount } =
+    partitionByCoreTitle(conceptRows);
+  const localIds = new Set(localNodes.map((node) => node.source_local_id));
+  return {
+    nodesToBackfill: localNodes.filter((node) =>
+      missingCoreTitleIds.has(node.source_local_id),
+    ),
+    withCoreTitleCount,
+    orphanedCount: difference(missingCoreTitleIds, localIds).size,
+  };
+};
+
+export const mergeNodesBySourceLocalId = (
+  nodes: RoamDiscourseNodeData[],
+  additionalNodes: RoamDiscourseNodeData[],
+): RoamDiscourseNodeData[] => {
+  const nodesById = new Map(nodes.map((node) => [node.source_local_id, node]));
+  for (const node of additionalNodes) {
+    if (!nodesById.has(node.source_local_id)) {
+      nodesById.set(node.source_local_id, node);
+    }
+  }
+  return [...nodesById.values()];
+};
diff --git a/apps/roam/src/utils/syncDgNodesToSupabase.ts b/apps/roam/src/utils/syncDgNodesToSupabase.ts
index d0c90168d..e7c8483be 100644
--- a/apps/roam/src/utils/syncDgNodesToSupabase.ts
+++ b/apps/roam/src/utils/syncDgNodesToSupabase.ts
@@ -26,6 +26,12 @@ import {
 } from "./convertRoamNodeToFullContent";
 import type { DGSupabaseClient } from "@repo/database/lib/client";
 import { intersection } from "@repo/utils/setOperations";
+import { CORE_TITLE_PROBE_SELECT } from "@repo/database/lib/coreTitleBackfill";
+import {
+  buildCoreTitleBackfill,
+  mergeNodesBySourceLocalId,
+  type CoreTitleBackfill,
+} from "./coreTitleBackfill";
 import type { Json, Enums } from "@repo/database/dbTypes";
 import { render as renderToast } from "roamjs-components/components/Toast";
 import internalError from "~/utils/internalError";
@@ -870,6 +876,44 @@ export const setSyncActivity = (active: boolean) => {
   }
 };
 
+const reportCoreTitleBackfill = ({
+  backfilled,
+  deferred,
+  skipped,
+  orphaned,
+}: {
+  backfilled: number;
+  deferred: number;
+  skipped: number;
+  orphaned: number;
+}): void => {
+  posthog.capture("Sync core_title backfill", {
+    backfilled,
+    deferred,
+    skipped,
+    orphaned,
+  });
+  if (backfilled === 0 && deferred === 0 && orphaned === 0) return;
+  const messages = [
+    `Backfilled core title for ${backfilled} node${backfilled === 1 ? "" : "s"}.`,
+    `${skipped} already had one.`,
+  ];
+  if (deferred > 0) {
+    messages.push(`${deferred} waiting for sync to be enabled.`);
+  }
+  if (orphaned > 0) {
+    messages.push(
+      `${orphaned} no longer match a discourse node in this graph.`,
+    );
+  }
+  renderToast({
+    id: "core-title-backfill",
+    intent: orphaned > 0 ? "warning" : "success",
+    content: messages.join(" "),
+    timeout: 5000,
+  });
+};
+
 const getAllMissingOrNewDiscourseNodes = async ({
   supabaseClient,
   spaceId,
@@ -880,9 +924,12 @@ const getAllMissingOrNewDiscourseNodes = async ({
   spaceId: number;
   since: number | undefined;
   nodeTypes: DiscourseNode[];
-}): Promise<RoamDiscourseNodeData[]> => {
+}): Promise<{
+  nodes: RoamDiscourseNodeData[];
+  coreTitleBackfill: CoreTitleBackfill | null;
+}> => {
   const allNodes = await getAllDiscourseNodesSince(undefined, nodeTypes);
-  if (since === undefined) return allNodes;
+  if (since === undefined) return { nodes: allNodes, coreTitleBackfill: null };
   const newNodes = await getAllDiscourseNodesSince(since, nodeTypes);
   const existingContentIdsReq = await getAllPages(
     supabaseClient
@@ -896,7 +943,7 @@ const getAllMissingOrNewDiscourseNodes = async ({
   const existingConceptIdsReq = await getAllPages(
     supabaseClient
       .from("my_concepts")
-      .select("source_local_id")
+      .select(CORE_TITLE_PROBE_SELECT)
       .eq("space_id", spaceId)
       .eq("arity", 0)
       .eq("is_schema", false)
@@ -911,10 +958,16 @@ const getAllMissingOrNewDiscourseNodes = async ({
     ),
     ...newNodes.map((n) => n.source_local_id),
   ]);
-  return [
-    ...newNodes,
-    ...allNodes.filter((n) => !existingIds.has(n.source_local_id)),
-  ];
+  return {
+    nodes: [
+      ...newNodes,
+      ...allNodes.filter((n) => !existingIds.has(n.source_local_id)),
+    ],
+    coreTitleBackfill: buildCoreTitleBackfill({
+      conceptRows: existingConceptIdsReq,
+      localNodes: allNodes,
+    }),
+  };
 };
 
 const getSharedNodeInstanceSourceLocalIds = async ({
@@ -1209,21 +1262,28 @@ export const createOrUpdateDiscourseEmbedding = async (
       (n) => n.backedBy === "user",
     );
 
-    const changedNodeInstances = await measureSyncPhase({
-      phase: isInitialSync
-        ? "getAllMissingOrNewDiscourseNodes"
-        : "getAllDiscourseNodesSince",
-      phases,
-      operation: () =>
-        isInitialSync
-          ? getAllMissingOrNewDiscourseNodes({
-              supabaseClient: activeSupabaseClient,
-              spaceId: activeContext.spaceId,
-              since: sinceTime,
-              nodeTypes: allDgNodeTypes,
-            })
-          : getAllDiscourseNodesSince(sinceTime, allDgNodeTypes),
-    });
+    const { nodes: changedNodeInstances, coreTitleBackfill } =
+      await measureSyncPhase({
+        phase: isInitialSync
+          ? "getAllMissingOrNewDiscourseNodes"
+          : "getAllDiscourseNodesSince",
+        phases,
+        operation: async () =>
+          isInitialSync
+            ? getAllMissingOrNewDiscourseNodes({
+                supabaseClient: activeSupabaseClient,
+                spaceId: activeContext.spaceId,
+                since: sinceTime,
+                nodeTypes: allDgNodeTypes,
+              })
+            : {
+                nodes: await getAllDiscourseNodesSince(
+                  sinceTime,
+                  allDgNodeTypes,
+                ),
+                coreTitleBackfill: null,
+              },
+      });
     const sharedSourceLocalIds = await measureSyncPhase({
       phase: "getSharedNodeInstanceSourceLocalIds",
       phases,
@@ -1238,6 +1298,16 @@ export const createOrUpdateDiscourseEmbedding = async (
           sharedSourceLocalIds.has(node.source_local_id),
         )
       : changedNodeInstances;
+    const nodesToBackfillCoreTitle = (
+      coreTitleBackfill?.nodesToBackfill ?? []
+    ).filter(
+      (node) =>
+        !sharedNodesOnlySync || sharedSourceLocalIds.has(node.source_local_id),
+    );
+    const conceptNodesToSync = mergeNodesBySourceLocalId(
+      nodeInstancesToSync,
+      nodesToBackfillCoreTitle,
+    );
     const sharedSourceLocalIdsToBackfill = await measureSyncPhase({
       phase: "getSharedSourceLocalIdsMissingFullContent",
       phases,
@@ -1316,7 +1386,7 @@ export const createOrUpdateDiscourseEmbedding = async (
       phases,
       operation: () =>
         convertDgToSupabaseConcepts({
-          nodesSince: nodeInstancesToSync,
+          nodesSince: conceptNodesToSync,
           since: sinceTime,
           allNodeTypes: allDgNodeTypes,
           sharedNodeTypeIds,
@@ -1324,6 +1394,16 @@ export const createOrUpdateDiscourseEmbedding = async (
           context: activeContext,
         }),
     });
+    if (coreTitleBackfill !== null) {
+      reportCoreTitleBackfill({
+        backfilled: nodesToBackfillCoreTitle.length,
+        deferred:
+          coreTitleBackfill.nodesToBackfill.length -
+          nodesToBackfillCoreTitle.length,
+        skipped: coreTitleBackfill.withCoreTitleCount,
+        orphaned: coreTitleBackfill.orphanedCount,
+      });
+    }
     await measureSyncPhase({
       phase: "cleanupOrphanedNodes",
       phases,
diff --git a/packages/database/src/lib/__tests__/coreTitleBackfill.test.ts b/packages/database/src/lib/__tests__/coreTitleBackfill.test.ts
new file mode 100644
index 000000000..6dfa6e33b
--- /dev/null
+++ b/packages/database/src/lib/__tests__/coreTitleBackfill.test.ts
@@ -0,0 +1,42 @@
+import { describe, expect, it } from "vitest";
+import { partitionByCoreTitle } from "../coreTitleBackfill";
+
+describe("partitionByCoreTitle", () => {
+  it("treats a null core_title as missing", () => {
+    const { missingCoreTitleIds, withCoreTitleCount } = partitionByCoreTitle([
+      { source_local_id: "a", core_title: null },
+      { source_local_id: "b", core_title: "already set" },
+    ]);
+
+    expect([...missingCoreTitleIds]).toEqual(["a"]);
+    expect(withCoreTitleCount).toBe(1);
+  });
+
+  it("counts an empty core_title as present", () => {
+    const { missingCoreTitleIds, withCoreTitleCount } = partitionByCoreTitle([
+      { source_local_id: "a", core_title: "" },
+    ]);
+
+    expect([...missingCoreTitleIds]).toEqual([]);
+    expect(withCoreTitleCount).toBe(1);
+  });
+
+  it("ignores rows without a source_local_id", () => {
+    const { missingCoreTitleIds, withCoreTitleCount } = partitionByCoreTitle([
+      { source_local_id: null, core_title: null },
+      { source_local_id: null, core_title: "set" },
+    ]);
+
+    expect([...missingCoreTitleIds]).toEqual([]);
+    expect(withCoreTitleCount).toBe(0);
+  });
+
+  it("returns nothing to backfill for an empty probe", () => {
+    const { missingCoreTitleIds, withCoreTitleCount } = partitionByCoreTitle(
+      [],
+    );
+
+    expect([...missingCoreTitleIds]).toEqual([]);
+    expect(withCoreTitleCount).toBe(0);
+  });
+});
diff --git a/packages/database/src/lib/coreTitleBackfill.ts b/packages/database/src/lib/coreTitleBackfill.ts
new file mode 100644
index 000000000..c563983dc
--- /dev/null
+++ b/packages/database/src/lib/coreTitleBackfill.ts
@@ -0,0 +1,23 @@
+export const CORE_TITLE_PROBE_SELECT =
+  "source_local_id, core_title:literal_content->>core_title";
+
+export type CoreTitleProbeRow = {
+  source_local_id: string | null;
+  core_title: string | null;
+};
+
+export const partitionByCoreTitle = (
+  rows: CoreTitleProbeRow[],
+): { missingCoreTitleIds: Set<string>; withCoreTitleCount: number } => {
+  const missingCoreTitleIds = new Set<string>();
+  let withCoreTitleCount = 0;
+  for (const row of rows) {
+    if (row.source_local_id === null) continue;
+    if (row.core_title === null) {
+      missingCoreTitleIds.add(row.source_local_id);
+    } else {
+      withCoreTitleCount += 1;
+    }
+  }
+  return { missingCoreTitleIds, withCoreTitleCount };
+};