diff --git a/app/javascript/mastodon/features/explore/components/story.jsx b/app/javascript/mastodon/features/explore/components/story.jsx
index 92eb41cff..80dd5200f 100644
--- a/app/javascript/mastodon/features/explore/components/story.jsx
+++ b/app/javascript/mastodon/features/explore/components/story.jsx
@@ -22,6 +22,7 @@ export default class Story extends PureComponent {
     author: PropTypes.string,
     sharedTimes: PropTypes.number,
     thumbnail: PropTypes.string,
+    thumbnailDescription: PropTypes.string,
     blurhash: PropTypes.string,
     expanded: PropTypes.bool,
   };
@@ -33,7 +34,7 @@ export default class Story extends PureComponent {
   handleImageLoad = () => this.setState({ thumbnailLoaded: true });
 
   render () {
-    const { expanded, url, title, lang, publisher, author, publishedAt, sharedTimes, thumbnail, blurhash } = this.props;
+    const { expanded, url, title, lang, publisher, author, publishedAt, sharedTimes, thumbnail, thumbnailDescription, blurhash } = this.props;
 
     const { thumbnailLoaded } = this.state;
 
@@ -49,7 +50,7 @@ export default class Story extends PureComponent {
           {thumbnail ? (
             <>
               <div className={classNames('story__thumbnail__preview', { 'story__thumbnail__preview--hidden': thumbnailLoaded })}><Blurhash hash={blurhash} /></div>
-              <img src={thumbnail} onLoad={this.handleImageLoad} alt='' role='presentation' />
+              <img src={thumbnail} onLoad={this.handleImageLoad} alt={thumbnailDescription} title={thumbnailDescription} lang={lang} />
             </>
           ) : <Skeleton />}
         </div>
diff --git a/app/javascript/mastodon/features/explore/links.jsx b/app/javascript/mastodon/features/explore/links.jsx
index 489ab6dd6..663aa6d80 100644
--- a/app/javascript/mastodon/features/explore/links.jsx
+++ b/app/javascript/mastodon/features/explore/links.jsx
@@ -67,6 +67,7 @@ class Links extends PureComponent {
             author={link.get('author_name')}
             sharedTimes={link.getIn(['history', 0, 'accounts']) * 1 + link.getIn(['history', 1, 'accounts']) * 1}
             thumbnail={link.get('image')}
+            thumbnailDescription={link.get('image_description')}
             blurhash={link.get('blurhash')}
           />
         ))}
diff --git a/app/javascript/mastodon/features/status/components/card.jsx b/app/javascript/mastodon/features/status/components/card.jsx
index 6ac3c1d0f..07fb1db9e 100644
--- a/app/javascript/mastodon/features/status/components/card.jsx
+++ b/app/javascript/mastodon/features/status/components/card.jsx
@@ -167,7 +167,8 @@ export default class Card extends PureComponent {
       />
     );
 
-    let thumbnail = <img src={card.get('image')} alt='' style={thumbnailStyle} onLoad={this.handleImageLoad} className='status-card__image-image' />;
+    const thumbnailDescription = card.get('image_description');
+    const thumbnail = <img src={card.get('image')} alt={thumbnailDescription} title={thumbnailDescription} lang={language} style={thumbnailStyle} onLoad={this.handleImageLoad} className='status-card__image-image' />;
 
     let spoilerButton = (
       <button type='button' onClick={this.handleReveal} className='spoiler-button__overlay'>
diff --git a/app/lib/link_details_extractor.rb b/app/lib/link_details_extractor.rb
index 4ee07acdc..b95ec8051 100644
--- a/app/lib/link_details_extractor.rb
+++ b/app/lib/link_details_extractor.rb
@@ -113,6 +113,7 @@ class LinkDetailsExtractor
       title: title || '',
       description: description || '',
       image_remote_url: image,
+      image_description: image_alt || '',
       type: type,
       link_type: link_type,
       width: width || 0,
@@ -168,6 +169,10 @@ class LinkDetailsExtractor
     valid_url_or_nil(opengraph_tag('og:image'))
   end
 
+  def image_alt
+    opengraph_tag('og:image:alt')
+  end
+
   def canonical_url
     valid_url_or_nil(link_tag('canonical') || opengraph_tag('og:url'), same_origin_only: true) || @original_url.to_s
   end
diff --git a/app/models/preview_card.rb b/app/models/preview_card.rb
index d9ddd3ba0..3e2b5bf99 100644
--- a/app/models/preview_card.rb
+++ b/app/models/preview_card.rb
@@ -31,6 +31,7 @@
 #  trendable                    :boolean
 #  link_type                    :integer
 #  published_at                 :datetime
+#  image_description            :string           default(""), not null
 #
 
 class PreviewCard < ApplicationRecord
diff --git a/app/serializers/rest/preview_card_serializer.rb b/app/serializers/rest/preview_card_serializer.rb
index cf700ff24..3e1c4bde3 100644
--- a/app/serializers/rest/preview_card_serializer.rb
+++ b/app/serializers/rest/preview_card_serializer.rb
@@ -6,7 +6,7 @@ class REST::PreviewCardSerializer < ActiveModel::Serializer
   attributes :url, :title, :description, :language, :type,
              :author_name, :author_url, :provider_name,
              :provider_url, :html, :width, :height,
-             :image, :embed_url, :blurhash, :published_at
+             :image, :image_description, :embed_url, :blurhash, :published_at
 
   def image
     object.image? ? full_asset_url(object.image.url(:original)) : nil
diff --git a/db/migrate/20230725213448_add_image_description_to_preview_cards.rb b/db/migrate/20230725213448_add_image_description_to_preview_cards.rb
new file mode 100644
index 000000000..03020db65
--- /dev/null
+++ b/db/migrate/20230725213448_add_image_description_to_preview_cards.rb
@@ -0,0 +1,17 @@
+# frozen_string_literal: true
+
+require Rails.root.join('lib', 'mastodon', 'migration_helpers')
+
+class AddImageDescriptionToPreviewCards < ActiveRecord::Migration[7.0]
+  include Mastodon::MigrationHelpers
+
+  disable_ddl_transaction!
+
+  def up
+    safety_assured { add_column_with_default :preview_cards, :image_description, :string, default: '', allow_null: false }
+  end
+
+  def down
+    remove_column :preview_cards, :image_description
+  end
+end
diff --git a/db/schema.rb b/db/schema.rb
index 0c8ede383..79bebb5ed 100644
--- a/db/schema.rb
+++ b/db/schema.rb
@@ -802,6 +802,7 @@ ActiveRecord::Schema[7.0].define(version: 2023_08_03_112520) do
     t.boolean "trendable"
     t.integer "link_type"
     t.datetime "published_at"
+    t.string "image_description", default: "", null: false
     t.index ["url"], name: "index_preview_cards_on_url", unique: true
   end
 
diff --git a/spec/lib/link_details_extractor_spec.rb b/spec/lib/link_details_extractor_spec.rb
index ef501efbf..599bc4e6d 100644
--- a/spec/lib/link_details_extractor_spec.rb
+++ b/spec/lib/link_details_extractor_spec.rb
@@ -3,33 +3,31 @@
 require 'rails_helper'
 
 RSpec.describe LinkDetailsExtractor do
-  subject { described_class.new(original_url, html, html_charset) }
+  subject { described_class.new(original_url, html, nil) }
 
-  let(:original_url) { '' }
-  let(:html) { '' }
-  let(:html_charset) { nil }
+  let(:original_url) { 'https://example.com/dog.html?tracking=123' }
 
   describe '#canonical_url' do
-    let(:original_url) { 'https://foo.com/article?bar=baz123' }
+    let(:html) { "<!doctype html><link rel='canonical' href='#{url}'>" }
+
+    context 'when canonical URL points to the same host' do
+      let(:url) { 'https://example.com/dog.html' }
+
+      it 'ignores the canonical URLs' do
+        expect(subject.canonical_url).to eq 'https://example.com/dog.html'
+      end
+    end
 
     context 'when canonical URL points to another host' do
-      let(:html) { '<!doctype html><link rel="canonical" href="https://bar.com/different-article" />' }
+      let(:url) { 'https://different.example.net/dog.html' }
 
       it 'ignores the canonical URLs' do
         expect(subject.canonical_url).to eq original_url
       end
     end
 
-    context 'when canonical URL points to the same host' do
-      let(:html) { '<!doctype html><link rel="canonical" href="https://foo.com/article" />' }
-
-      it 'ignores the canonical URLs' do
-        expect(subject.canonical_url).to eq 'https://foo.com/article'
-      end
-    end
-
     context 'when canonical URL is set to "null"' do
-      let(:html) { '<!doctype html><link rel="canonical" href="null" />' }
+      let(:url) { 'null' }
 
       it 'ignores the canonical URLs' do
         expect(subject.canonical_url).to eq original_url
@@ -37,46 +35,103 @@ RSpec.describe LinkDetailsExtractor do
     end
   end
 
+  context 'when only basic metadata is present' do
+    let(:html) { <<~HTML }
+      <!doctype html>
+      <html lang="en">
+      <head>
+        <title>Man bites dog</title>
+        <meta name="description" content="A dog&#39;s tale">
+      </head>
+      </html>
+    HTML
+
+    describe '#title' do
+      it 'returns the title from title tag' do
+        expect(subject.title).to eq 'Man bites dog'
+      end
+    end
+
+    describe '#description' do
+      it 'returns the description from meta tag' do
+        expect(subject.description).to eq "A dog's tale"
+      end
+    end
+
+    describe '#language' do
+      it 'returns the language from lang attribute' do
+        expect(subject.language).to eq 'en'
+      end
+    end
+  end
+
   context 'when structured data is present' do
-    let(:original_url) { 'https://example.com/page.html' }
-
-    context 'when is wrapped in CDATA tags' do
-      let(:html) { <<~HTML }
-        <!doctype html>
-        <html>
-        <head>
-          <script type="application/ld+json">
-          //<![CDATA[
-          {"@context":"http://schema.org","@type":"NewsArticle","mainEntityOfPage":"https://example.com/page.html","headline":"Foo","datePublished":"2022-01-31T19:53:00+00:00","url":"https://example.com/page.html","description":"Bar","author":{"@type":"Person","name":"Hoge"},"publisher":{"@type":"Organization","name":"Baz"}}
-          //]]>
-          </script>
-        </head>
-        </html>
-      HTML
+    let(:ld_json) do
+      {
+        '@context' => 'https://schema.org',
+        '@type' => 'NewsArticle',
+        'headline' => 'Man bites dog',
+        'description' => "A dog's tale",
+        'datePublished' => '2022-01-31T19:53:00+00:00',
+        'author' => {
+          '@type' => 'Organization',
+          'name' => 'Charlie Brown',
+        },
+        'publisher' => {
+          '@type' => 'NewsMediaOrganization',
+          'name' => 'Pet News',
+          'url' => 'https://example.com',
+        },
+      }.to_json
+    end
 
+    shared_examples 'structured data' do
       describe '#title' do
         it 'returns the title from structured data' do
-          expect(subject.title).to eq 'Foo'
+          expect(subject.title).to eq 'Man bites dog'
         end
       end
 
       describe '#description' do
         it 'returns the description from structured data' do
-          expect(subject.description).to eq 'Bar'
+          expect(subject.description).to eq "A dog's tale"
         end
       end
 
-      describe '#provider_name' do
-        it 'returns the provider name from structured data' do
-          expect(subject.provider_name).to eq 'Baz'
+      describe '#published_at' do
+        it 'returns the publicaton time from structured data' do
+          expect(subject.published_at).to eq '2022-01-31T19:53:00+00:00'
         end
       end
 
       describe '#author_name' do
         it 'returns the author name from structured data' do
-          expect(subject.author_name).to eq 'Hoge'
+          expect(subject.author_name).to eq 'Charlie Brown'
         end
       end
+
+      describe '#provider_name' do
+        it 'returns the provider name from structured data' do
+          expect(subject.provider_name).to eq 'Pet News'
+        end
+      end
+    end
+
+    context 'when is wrapped in CDATA tags' do
+      let(:html) { <<~HTML }
+        <!doctype html>
+        <html>
+          <head>
+            <script type="application/ld+json">
+              //<![CDATA[
+              #{ld_json}
+              //]]>
+            </script>
+          </head>
+        </html>
+      HTML
+
+      include_examples 'structured data'
     end
 
     context 'with the first tag is invalid JSON' do
@@ -85,76 +140,152 @@ RSpec.describe LinkDetailsExtractor do
         <html>
         <body>
           <script type="application/ld+json">
-            {
-              "@context":"https://schema.org",
-              "@type":"ItemList",
-              "url":"https://example.com/page.html",
-              "name":"Foo",
-              "description":"Bar"
-            },
-            {
-              "@context": "https://schema.org",
-              "@type": "BreadcrumbList",
-              "itemListElement":[
-                {
-                  "@type":"ListItem",
-                  "position":1,
-                  "item":{
-                    "@id":"https://www.example.com",
-                    "name":"Baz"
-                  }
-                }
-              ]
-            }
+            invalid LD+JSON
           </script>
           <script type="application/ld+json">
-            {
-              "@context":"https://schema.org",
-              "@type":"NewsArticle",
-              "mainEntityOfPage": {
-                "@type":"WebPage",
-                "@id": "http://example.com/page.html"
-              },
-              "headline": "Foo",
-              "description": "Bar",
-              "datePublished": "2022-01-31T19:46:00+00:00",
-              "author": {
-                "@type": "Organization",
-                "name": "Hoge"
-              },
-              "publisher": {
-                "@type": "NewsMediaOrganization",
-                "name":"Baz",
-                "url":"https://example.com/"
-              }
-            }
+            #{ld_json}
           </script>
         </body>
         </html>
       HTML
 
-      describe '#title' do
-        it 'returns the title from structured data' do
-          expect(subject.title).to eq 'Foo'
-        end
-      end
+      include_examples 'structured data'
+    end
 
-      describe '#description' do
-        it 'returns the description from structured data' do
-          expect(subject.description).to eq 'Bar'
-        end
-      end
+    context 'with preceding block of unsupported LD+JSON' do
+      let(:html) { <<~HTML }
+        <!doctype html>
+        <html>
+        <body>
+          <script type="application/ld+json">
+            [
+              {
+                "@context": "https://schema.org",
+                "@type": "ItemList",
+                "url": "https://example.com/cat.html",
+                "name": "Man bites cat",
+                "description": "A cat's tale"
+              },
+              {
+                "@context": "https://schema.org",
+                "@type": "BreadcrumbList",
+                "itemListElement":[
+                  {
+                    "@type": "ListItem",
+                    "position": 1,
+                    "item": {
+                      "@id": "https://www.example.com",
+                      "name": "Cat News"
+                    }
+                  }
+                ]
+              }
+            ]
+          </script>
+          <script type="application/ld+json">
+            #{ld_json}
+          </script>
+        </body>
+        </html>
+      HTML
 
-      describe '#provider_name' do
-        it 'returns the provider name from structured data' do
-          expect(subject.provider_name).to eq 'Baz'
-        end
-      end
+      include_examples 'structured data'
+    end
 
-      describe '#author_name' do
-        it 'returns the author name from structured data' do
-          expect(subject.author_name).to eq 'Hoge'
-        end
+    context 'with unsupported in same block LD+JSON' do
+      let(:html) { <<~HTML }
+        <!doctype html>
+        <html>
+        <body>
+          <script type="application/ld+json">
+            [
+              {
+                "@context": "https://schema.org",
+                "@type": "ItemList",
+                "url": "https://example.com/cat.html",
+                "name": "Man bites cat",
+                "description": "A cat's tale"
+              },
+              #{ld_json}
+            ]
+          </script>
+        </body>
+        </html>
+      HTML
+
+      include_examples 'structured data'
+    end
+  end
+
+  context 'when Open Graph protocol data is present' do
+    let(:html) { <<~HTML }
+      <!doctype html>
+      <html>
+      <head>
+        <meta property="og:url" content="https://example.com/dog.html">
+        <meta property="og:title" content="Man bites dog">
+        <meta property="og:description" content="A dog's tale">
+        <meta property="article:published_time" content="2022-01-31T19:53:00+00:00">
+        <meta property="og:author" content="Charlie Brown">
+        <meta property="og:locale" content="en">
+        <meta property="og:image" content="https://example.com/snoopy.jpg">
+        <meta property="og:image:alt" content="A good boy">
+        <meta property="og:site_name" content="Pet News">
+      </head>
+      </html>
+    HTML
+
+    describe '#canonical_url' do
+      it 'returns the URL from Open Graph protocol data' do
+        expect(subject.canonical_url).to eq 'https://example.com/dog.html'
+      end
+    end
+
+    describe '#title' do
+      it 'returns the title from Open Graph protocol data' do
+        expect(subject.title).to eq 'Man bites dog'
+      end
+    end
+
+    describe '#description' do
+      it 'returns the description from Open Graph protocol data' do
+        expect(subject.description).to eq "A dog's tale"
+      end
+    end
+
+    describe '#published_at' do
+      it 'returns the publicaton time from Open Graph protocol data' do
+        expect(subject.published_at).to eq '2022-01-31T19:53:00+00:00'
+      end
+    end
+
+    describe '#author_name' do
+      it 'returns the author name from Open Graph protocol data' do
+        expect(subject.author_name).to eq 'Charlie Brown'
+      end
+    end
+
+    describe '#language' do
+      it 'returns the language from Open Graph protocol data' do
+        expect(subject.language).to eq 'en'
+      end
+    end
+
+    describe '#image' do
+      it 'returns the image from Open Graph protocol data' do
+        expect(subject.image).to eq 'https://example.com/snoopy.jpg'
+      end
+    end
+
+    describe '#image:alt' do
+      it 'returns the image description from Open Graph protocol data' do
+        expect(subject.image_alt).to eq 'A good boy'
+      end
+    end
+
+    describe '#provider_name' do
+      it 'returns the provider name from Open Graph protocol data' do
+        expect(subject.provider_name).to eq 'Pet News'
       end
     end
   end