diff --git a/README.md b/README.md
index d0267fa..9d26f54 100644
--- a/README.md
+++ b/README.md
@@ -19,12 +19,13 @@ And here are the current Ghost events / webhooks and their endpoints:
 | Post unpublished       | .../ghost/delete-record        |
 | Post deleted           | .../ghost/delete-record        |
 
-Finally, here are the currently configured Algolia indices for each publication:
+Finally, here are the currently configured Algolia indices for each publication, which are all production unless otherwise noted:
 
 | Publication URL                               | CMS      | Algolia index |
 | --------------------------------------------- | -------- | ------------- |
 | https://www.freecodecamp.org/news/            | Hashnode | news          |
-| https://chinese.freecodecamp.org/news/        | Ghost    | news-zh       |
+| http://localhost:3030                         | Ghost    | news-es (dev) |
+| https://www.freecodecamp.org/chinese/news/    | Ghost    | news-zh       |
 | https://www.freecodecamp.org/espanol/news/    | Ghost    | news-es       |
 | https://www.freecodecamp.org/italian/news/    | Ghost    | news-it       |
 | https://www.freecodecamp.org/japanese/news/   | Ghost    | news-ja       |
diff --git a/lib/utils/helpers.js b/lib/utils/helpers.js
index 8abcd04..59bca3b 100644
--- a/lib/utils/helpers.js
+++ b/lib/utils/helpers.js
@@ -6,9 +6,14 @@ export const algoliaClient = algoliasearch(
   process.env.ALGOLIA_ADMIN_KEY
 );
 
-export const searchIndexNameMap = {
-  'http://localhost:3030': 'news-es', // Dockerized Spanish Ghost instance for testing, which is mimicking the old English instance
-  'https://chinese.freecodecamp.org/news/': 'news-zh',
+const searchIndexNameMap = {
+  // Dockerized Spanish Ghost instance for testing
+  'http://localhost:3030': 'news-es',
+  // The Chinese publication is at https://chinese.freecodecamp.org/news/,
+  // but this map and the getSearchIndexName function are used after formatting
+  // the post for Algolia, which updates necessary URLs for the search record
+  // to start with https://www.freecodecamp.org/chinese/news/
+  'https://www.freecodecamp.org/chinese/news/': 'news-zh',
   'https://www.freecodecamp.org/espanol/news/': 'news-es',
   'https://www.freecodecamp.org/italian/news/': 'news-it',
   'https://www.freecodecamp.org/japanese/news/': 'news-ja',
@@ -96,6 +101,30 @@ export const formatHashnodePost = (post) => {
   };
 };
 
+export const getBaseSiteURL = (url) => {
+  const URLObj = new URL(url);
+  const { host, pathname } = URLObj;
+  const pathParts = pathname.split('/').filter(Boolean);
+  let siteLang;
+  if (host.startsWith('chinese')) {
+    siteLang = 'chinese';
+  } else if (pathParts.length === 3) {
+    // Webhooks will only be triggered by posts, so the path will
+    // always have 3 parts for our localized instances:
+    // (/<lang>/news/<slug>/).
+    // Or if it's coming from a Dockerized test instance / localhost,
+    // it will only have 1 part: (/<slug>/).
+    siteLang = pathParts[0];
+  }
+  const computedPath = siteLang ? `/${siteLang}/news/` : '/news/';
+
+  if (host.startsWith('localhost')) {
+    return `http://${host}${computedPath}`;
+  } else {
+    return `https://www.freecodecamp.org${computedPath}`;
+  }
+};
+
 export const formatGhostPost = (post) => {
   const {
     id,
@@ -107,22 +136,7 @@ export const formatGhostPost = (post) => {
     feature_image,
     published_at
   } = post;
-  const URLObj = new URL(url);
-  const { href, origin, pathname } = URLObj;
-  const pathParts = pathname.split('/').filter(Boolean);
-  let siteLang;
-  if (href.startsWith('https://chinese.freecodecamp.org/')) {
-    siteLang = 'chinese';
-  } else if (pathParts.length === 3) {
-    // Webhooks will only be triggered by posts, so the path will
-    // always have 3 parts for our localized instances:
-    // (/<lang>/news/<slug>/).
-    // Or if it's coming from a Dockerized test instance, it will
-    // only have 1 part: (/<slug>/).
-    siteLang = pathParts[0];
-  }
-  const siteURL = `${origin}/${siteLang ? `${siteLang}/` : ''}news/`;
-  console.log({ siteURL, siteLang, href, origin, pathname, pathParts });
+  const siteURL = getBaseSiteURL(url);
 
   return {
     objectID: id,
diff --git a/packages/ghost/delete-record/index.js b/packages/ghost/delete-record/index.js
index 69261ab..19b8427 100644
--- a/packages/ghost/delete-record/index.js
+++ b/packages/ghost/delete-record/index.js
@@ -1,6 +1,10 @@
 // The ../../../lib/utils directory is zipped in the same directory
 // as the function during the build process
-import { algoliaClient, getSearchIndexName } from './utils/helpers.js';
+import {
+  algoliaClient,
+  getBaseSiteURL,
+  getSearchIndexName
+} from './utils/helpers.js';
 
 export const deleteRecord = async (req) => {
   try {
@@ -15,13 +19,16 @@ export const deleteRecord = async (req) => {
     // Whether a published post is unpublished or deleted, the
     // status will be 'published' in the previous state
     if (prevState.status === 'published') {
-      // Deleted posts don't include a url. But since every
-      // post must include at least one author, set the index
-      // based on the primary author page url instead
-      const primaryAuthorUrl = prevState.authors
+      // Deleted posts don't include a url or a primary author object.
+      // But since every post returns an author array, we can use
+      // the first author object to determine the search index name.
+      // This helps since we're handling data from localhost,
+      // chinese.freecodecamp.org, and www.freecodecamp.org.
+      const primaryAuthorURL = prevState.authors
         ? prevState.authors[0].url
         : currState.authors[0].url;
-      const indexName = getSearchIndexName(primaryAuthorUrl);
+      const siteURL = getBaseSiteURL(primaryAuthorURL);
+      const indexName = getSearchIndexName(siteURL);
 
       if (!indexName) {
         throw new Error('No matching index found for the current post');