diff --git a/.gitignore b/.gitignore
index 72dfb7711..b9dddd528 100644
--- a/.gitignore
+++ b/.gitignore
@@ -19,3 +19,6 @@ npm-debug.log*
 yarn-debug.log*
 yarn-error.log*
 .aider*
+
+# Temporary exclusion for Claude project 4/10/26
+# md-sentence-case*
\ No newline at end of file
diff --git a/docs/endpoints/md-sentence-case-run.bat b/docs/endpoints/md-sentence-case-run.bat
new file mode 100644
index 000000000..f0263addf
--- /dev/null
+++ b/docs/endpoints/md-sentence-case-run.bat
@@ -0,0 +1 @@
+node md-sentence-case.js
\ No newline at end of file
diff --git a/docs/endpoints/md-sentence-case.js b/docs/endpoints/md-sentence-case.js
new file mode 100644
index 000000000..063b33f8d
--- /dev/null
+++ b/docs/endpoints/md-sentence-case.js
@@ -0,0 +1,206 @@
+#!/usr/bin/env node
+
+/**
+ * md-sentence-case.js
+ *
+ * Converts Markdown headings (h2–h4) and heading link display text
+ * from title case to sentence case, with a configurable exclusions list.
+ *
+ * Usage:
+ *   node md-sentence-case.js [--dry-run] [directory]
+ *
+ * Options:
+ *   --dry-run    Preview changes without writing any files
+ *   directory    Path to the repo/folder to process (defaults to current dir)
+ *
+ * Examples:
+ *   node md-sentence-case.js --dry-run ./docs
+ *   node md-sentence-case.js ./docs
+ *   node md-sentence-case.js            # runs on current directory
+ */
+
+const fs   = require('fs');
+const path = require('path');
+
+// ─── CONFIGURE YOUR EXCLUSIONS HERE ──────────────────────────────────────────
+//
+// Words and phrases listed here will NOT be lowercased.
+// - Matching is case-insensitive, so "aws" and "AWS" are both caught.
+// - Put longer/more specific phrases BEFORE shorter ones so they match first
+//   (e.g. "Amazon Web Services" before "AWS").
+
+const EXCLUSIONS = [
+  // ── Multi-word phrases first (must come before single words they contain) ──
+  'Mobile SDK',
+  'Amazon S3',
+  'Secure Signals',
+  'Prebid.js',
+  'FAQs',
+
+  // ── Acronyms & initialisms ─────────────────────────────────────────────────
+  'API',
+  'AWS',
+  'CTV',
+  'DII',
+  'DSP',
+  'EUID',
+  'FAQ',
+  'GMA',
+  'ID',
+  'IMA',
+  'JSON',
+  'POST',
+  'SDK',
+  'SSO',
+  'UID2',
+  'URL',
+
+  // ── Proper nouns & brand names ─────────────────────────────────────────────
+  'Google',
+  'Gradle',
+  'GraphQL',
+  'I',           // first-person pronoun
+  'JavaScript',
+  'Maven',
+  'Prebid',
+  'ProGuard',
+  'Sharing',
+  'Swift',
+  'Base64',
+  'SHA-256',
+
+  // ── Add your own exclusions below this line ────────────────────────────────
+
+];
+
+// ─────────────────────────────────────────────────────────────────────────────
+
+const args    = process.argv.slice(2);
+const dryRun  = args.includes('--dry-run');
+const target  = args.find(a => !a.startsWith('--')) || '.';
+
+/**
+ * Escapes a string for safe use inside a RegExp.
+ */
+function escapeForRegex(str) {
+  return str.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+}
+
+/**
+ * Converts a heading/link text string to sentence case,
+ * then restores any words in the EXCLUSIONS list.
+ */
+function toSentenceCase(text) {
+  // Lowercase everything, then uppercase the very first character.
+  let result = text.charAt(0).toUpperCase() + text.slice(1).toLowerCase();
+
+  // Restore excluded words. Longer phrases are matched first (list order).
+  for (const word of EXCLUSIONS) {
+    const pattern = new RegExp(`\\b${escapeForRegex(word)}\\b`, 'gi');
+    result = result.replace(pattern, word);
+  }
+
+  return result;
+}
+
+/**
+ * Processes the content of a single Markdown file.
+ * Returns the updated content and a flag indicating whether anything changed.
+ */
+function processContent(content, filePath) {
+  let changed = false;
+  const relPath = path.relative(process.cwd(), filePath);
+
+  // Detect and normalise line endings so the heading regex works on all
+  // platforms. Windows files use \r\n; the \r is a line terminator in JS
+  // regex, which causes (.+) to stop short and the heading match to fail.
+  const lineEnding = content.includes('\r\n') ? '\r\n' : '\n';
+  const lines = content.replace(/\r\n/g, '\n').replace(/\r/g, '\n').split('\n');
+
+  const processed = lines.map((line, idx) => {
+    const lineNum = idx + 1;
+    let newLine = line;
+
+    // ── 1. Headings h2–h4 ──────────────────────────────────────────────────
+    newLine = newLine.replace(/^(#{2,4} )(.+?)(\s*)$/, (_match, prefix, headingText, trailing) => {
+      const converted = toSentenceCase(headingText);
+      if (converted !== headingText) {
+        console.log(`  ${relPath}:${lineNum}  heading  "${headingText}"  →  "${converted}"`);
+        changed = true;
+      }
+      return prefix + converted + trailing;
+    });
+
+    // ── 2. Display text of internal/relative links ────────────────────────
+    // Matches [Text](url) where url is NOT an external http/https link,
+    // covering anchor links (#...), relative file paths (../foo.md), etc.
+    newLine = newLine.replace(/\[([^\]]+)\]\(((?!https?:\/\/)[^)]+)\)/g, (_match, displayText, url) => {
+      const converted = toSentenceCase(displayText);
+      if (converted !== displayText) {
+        console.log(`  ${relPath}:${lineNum}  link     "[${displayText}]"  →  "[${converted}]"`);
+        changed = true;
+      }
+      return `[${converted}](${url})`;
+    });
+
+    return newLine;
+  });
+
+  // Restore the original line endings when reassembling the file.
+  return { content: processed.join(lineEnding), changed };
+}
+
+/**
+ * Recursively finds all .md files under a directory,
+ * skipping node_modules and .git.
+ */
+function findMarkdownFiles(dir) {
+  const results = [];
+
+  for (const entry of fs.readdirSync(dir, { withFileTypes: true })) {
+    const fullPath = path.join(dir, entry.name);
+
+    if (entry.isDirectory()) {
+      if (entry.name === 'node_modules' || entry.name === '.git') continue;
+      results.push(...findMarkdownFiles(fullPath));
+    } else if (entry.isFile() && /\.mdx?$/.test(entry.name)) {
+      results.push(fullPath);
+    }
+  }
+
+  return results;
+}
+
+// ─── MAIN ─────────────────────────────────────────────────────────────────────
+
+const absTarget = path.resolve(target);
+const files     = findMarkdownFiles(absTarget);
+
+console.log(`\nmd-sentence-case${dryRun ? ' [DRY RUN]' : ''}`);
+console.log(`Target : ${absTarget}`);
+console.log(`Files  : ${files.length} Markdown file(s) found\n`);
+
+if (files.length === 0) {
+  console.log('Nothing to do.');
+  process.exit(0);
+}
+
+let totalChanged = 0;
+
+for (const file of files) {
+  const original              = fs.readFileSync(file, 'utf8');
+  const { content, changed }  = processContent(original, file);
+
+  if (changed) {
+    totalChanged++;
+    if (!dryRun) {
+      fs.writeFileSync(file, content, 'utf8');
+    }
+  }
+}
+
+const verb = dryRun ? 'Would modify' : 'Modified';
+console.log(`\n${verb} ${totalChanged} of ${files.length} file(s).`);
+if (dryRun) {
+  console.log('Run without --dry-run to apply changes.');
+}