diff --git a/lib/rename-version.js b/lib/rename-version.js
index 4311539f6c..d0bb765957 100755
--- a/lib/rename-version.js
+++ b/lib/rename-version.js
@@ -12,6 +12,7 @@ const fs = require('fs');
 const glob = require('glob');
 const path = require('path');
 const readMetadata = require('./server/readMetadata.js');
+const metadataUtils = require('./server/metadataUtils.js');
 
 const CWD = process.cwd();
 
@@ -99,7 +100,7 @@ if (fs.existsSync(CWD + '/versioned_docs/version-' + currentVersion)) {
     if (extension !== '.md' && extension !== '.markdown') {
       return;
     }
-    const res = readMetadata.extractMetadata(fs.readFileSync(file, 'utf8'));
+    const res = metadataUtils.extractMetadata(fs.readFileSync(file, 'utf8'));
     const metadata = res.metadata;
     const rawContent = res.rawContent;
     if (!metadata.id) {
diff --git a/lib/server/generate.js b/lib/server/generate.js
index 776cabff84..dfe008e95e 100644
--- a/lib/server/generate.js
+++ b/lib/server/generate.js
@@ -8,6 +8,8 @@
 async function execute() {
   const extractTranslations = require('../write-translations.js');
 
+  const metadataUtils = require('./metadataUtils');
+
   const CWD = process.cwd();
   const cssnano = require('cssnano');
   const fs = require('fs-extra');
@@ -147,8 +149,9 @@ async function execute() {
       return;
     }
 
-    let rawContent = readMetadata.extractMetadata(fs.readFileSync(file, 'utf8'))
-      .rawContent;
+    let rawContent = metadataUtils.extractMetadata(
+      fs.readFileSync(file, 'utf8')
+    ).rawContent;
 
     const language = metadata.language;
 
@@ -261,7 +264,7 @@ async function execute() {
         .replace('-', '/')
         .replace('-', '/')
         .replace(/\.md$/, '.html');
-      const result = readMetadata.extractMetadata(
+      const result = metadataUtils.extractMetadata(
         fs.readFileSync(normalizedFile, {encoding: 'utf8'})
       );
       const rawContent = result.rawContent;
diff --git a/lib/server/metadataUtils.js b/lib/server/metadataUtils.js
new file mode 100644
index 0000000000..16f801f14e
--- /dev/null
+++ b/lib/server/metadataUtils.js
@@ -0,0 +1,66 @@
+/**
+ * Copyright (c) 2017-present, Facebook, Inc.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+// split markdown header
+function splitHeader(content) {
+  // New line characters need to handle all operating systems.
+  const lines = content.split(/\r?\n/);
+  if (lines[0] !== '---') {
+    return {};
+  }
+  let i = 1;
+  for (; i < lines.length - 1; ++i) {
+    if (lines[i] === '---') {
+      break;
+    }
+  }
+  return {
+    header: lines.slice(1, i + 1).join('\n'),
+    content: lines.slice(i + 1).join('\n'),
+  };
+}
+
+// Extract markdown metadata header
+function extractMetadata(content) {
+  const metadata = {};
+  const both = splitHeader(content);
+
+  // if no content returned, then that means there was no header, and both.header is the content
+  if (!both.content) {
+    if (!both.header) {
+      // if no both returned, then that means there was no header and no content => we return the current content of the file
+      return {metadata, rawContent: content};
+    }
+    return {metadata, rawContent: both.header};
+  }
+
+  // New line characters => to handle all operating systems.
+  const lines = both.header.split(/\r?\n/);
+
+  // Loop that add to metadata the current content of the fields of the header
+  // Like the format:
+  // id:
+  // title:
+  // original_id:
+  for (let i = 0; i < lines.length - 1; ++i) {
+    const keyvalue = lines[i].split(':');
+    const key = keyvalue[0].trim();
+    let value = keyvalue
+      .slice(1)
+      .join(':')
+      .trim();
+    try {
+      value = JSON.parse(value);
+    } catch (e) {}
+    metadata[key] = value;
+  }
+  return {metadata, rawContent: both.content};
+}
+
+module.exports = {
+  extractMetadata,
+};
diff --git a/lib/server/readMetadata.js b/lib/server/readMetadata.js
index 21677019da..760b2903cb 100644
--- a/lib/server/readMetadata.js
+++ b/lib/server/readMetadata.js
@@ -12,6 +12,8 @@ const fs = require('fs');
 const glob = require('glob');
 const chalk = require('chalk');
 
+const metadataUtils = require('./metadataUtils');
+
 const env = require('./env.js');
 const siteConfig = require(CWD + '/siteConfig.js');
 const versionFallback = require('./versionFallback.js');
@@ -79,52 +81,11 @@ function readSidebar() {
   return order;
 }
 
-// split markdown header
-function splitHeader(content) {
-  // New line characters need to handle all operating systems.
-  const lines = content.split(/\r?\n/);
-  if (lines[0] !== '---') {
-    return {};
-  }
-  let i = 1;
-  for (; i < lines.length - 1; ++i) {
-    if (lines[i] === '---') {
-      break;
-    }
-  }
-  return {
-    header: lines.slice(1, i + 1).join('\n'),
-    content: lines.slice(i + 1).join('\n'),
-  };
-}
-
-// Extract markdown metadata header
-function extractMetadata(content) {
-  const metadata = {};
-  const both = splitHeader(content);
-  if (Object.keys(both).length === 0) {
-    return {metadata, rawContent: content};
-  }
-  const lines = both.header.split('\n');
-  for (let i = 0; i < lines.length - 1; ++i) {
-    const keyvalue = lines[i].split(':');
-    const key = keyvalue[0].trim();
-    let value = keyvalue
-      .slice(1)
-      .join(':')
-      .trim();
-    try {
-      value = JSON.parse(value);
-    } catch (e) {}
-    metadata[key] = value;
-  }
-  return {metadata, rawContent: both.content};
-}
-
 // process the metadata for a document found in either 'docs' or 'translated_docs'
 function processMetadata(file, refDir) {
-  const result = extractMetadata(fs.readFileSync(file, 'utf8'));
+  const result = metadataUtils.extractMetadata(fs.readFileSync(file, 'utf8'));
   const language = utils.getLanguage(file, refDir) || 'en';
+
   const metadata = {};
   for (const fieldName of Object.keys(result.metadata)) {
     if (SupportedHeaderFields.has(fieldName)) {
@@ -367,7 +328,9 @@ function generateMetadataBlog() {
         .replace('-', '/')
         .replace('-', '/')
         .replace(/\.md$/, '.html');
-      const result = extractMetadata(fs.readFileSync(file, {encoding: 'utf8'}));
+      const result = metadataUtils.extractMetadata(
+        fs.readFileSync(file, {encoding: 'utf8'})
+      );
       const rawContent = result.rawContent;
       const metadata = Object.assign(
         {path: filePath, content: rawContent},
@@ -414,7 +377,6 @@ function generateMetadataBlog() {
 module.exports = {
   getDocsPath,
   readSidebar,
-  extractMetadata,
   processMetadata,
   generateMetadataDocs,
   generateMetadataBlog,
diff --git a/lib/server/server.js b/lib/server/server.js
index be1c342a46..3e5ac7bdc4 100644
--- a/lib/server/server.js
+++ b/lib/server/server.js
@@ -8,6 +8,8 @@
 function execute(port) {
   const extractTranslations = require('../write-translations');
 
+  const metadataUtils = require('./metadataUtils');
+
   const env = require('./env.js');
   const translation = require('./translation');
   const express = require('express');
@@ -186,8 +188,9 @@ function execute(port) {
       return;
     }
 
-    let rawContent = readMetadata.extractMetadata(fs.readFileSync(file, 'utf8'))
-      .rawContent;
+    let rawContent = metadataUtils.extractMetadata(
+      fs.readFileSync(file, 'utf8')
+    ).rawContent;
 
     // generate table of contents if appropriate
     if (rawContent && rawContent.indexOf(TABLE_OF_CONTENTS_TOKEN) !== -1) {
@@ -326,7 +329,7 @@ function execute(port) {
       file = file.replace(new RegExp('/', 'g'), '-');
       file = join(CWD, 'blog', file);
 
-      const result = readMetadata.extractMetadata(
+      const result = metadataUtils.extractMetadata(
         fs.readFileSync(file, {encoding: 'utf8'})
       );
       let rawContent = result.rawContent;
diff --git a/lib/server/versionFallback.js b/lib/server/versionFallback.js
index 85c3a9394a..1b5f5a9a55 100644
--- a/lib/server/versionFallback.js
+++ b/lib/server/versionFallback.js
@@ -11,6 +11,8 @@ const fs = require('fs');
 const path = require('path');
 const assert = require('assert');
 
+const metadataUtils = require('./metadataUtils');
+
 const env = require('./env.js');
 const utils = require('./utils.js');
 const siteConfig = require(CWD + '/siteConfig.js');
@@ -37,50 +39,6 @@ if (fs.existsSync(CWD + '/languages.js')) {
   ];
 }
 
-/*****************************************************************/
-
-// included to prevent cyclical dependency with readMetadata.js
-
-function splitHeader(content) {
-  const lines = content.split(/\r?\n/);
-  let i = 1;
-  for (; i < lines.length - 1; ++i) {
-    if (lines[i] === '---') {
-      break;
-    }
-  }
-  return {
-    header: lines.slice(1, i + 1).join('\n'),
-    content: lines.slice(i + 1).join('\n'),
-  };
-}
-
-// Extract markdown metadata header
-function extractMetadata(content) {
-  const metadata = {};
-  const both = splitHeader(content);
-  // if no content returned, then that means there was no header, and both.header is the content
-  if (!both.content) {
-    return {metadata, rawContent: both.header};
-  }
-  const lines = both.header.split(/\r?\n/);
-  for (let i = 0; i < lines.length - 1; ++i) {
-    const keyvalue = lines[i].split(':');
-    const key = keyvalue[0].trim();
-    let value = keyvalue
-      .slice(1)
-      .join(':')
-      .trim();
-    try {
-      value = JSON.parse(value);
-    } catch (e) {}
-    metadata[key] = value;
-  }
-  return {metadata, rawContent: both.content};
-}
-
-/*****************************************************************/
-
 const versionFolder = CWD + '/versioned_docs/';
 
 // available stores doc ids of documents that are available for
@@ -95,7 +53,7 @@ files.forEach(file => {
   if (ext !== '.md' && ext !== '.markdown') {
     return;
   }
-  const res = extractMetadata(fs.readFileSync(file, 'utf8'));
+  const res = metadataUtils.extractMetadata(fs.readFileSync(file, 'utf8'));
   const metadata = res.metadata;
 
   if (!metadata.original_id) {
@@ -186,15 +144,20 @@ function diffLatestDoc(file, id) {
   }
 
   return (
-    extractMetadata(fs.readFileSync(latestFile, 'utf8')).rawContent.trim() !==
-    extractMetadata(fs.readFileSync(file, 'utf8')).rawContent.trim()
+    metadataUtils
+      .extractMetadata(fs.readFileSync(latestFile, 'utf8'))
+      .rawContent.trim() !==
+    metadataUtils
+      .extractMetadata(fs.readFileSync(file, 'utf8'))
+      .rawContent.trim()
   );
 }
 
 // return metadata for a versioned file given the file, its version (requested),
 // the version of the file to be used, and its language
 function processVersionMetadata(file, version, useVersion, language) {
-  const metadata = extractMetadata(fs.readFileSync(file, 'utf8')).metadata;
+  const metadata = metadataUtils.extractMetadata(fs.readFileSync(file, 'utf8'))
+    .metadata;
 
   // Add subdirectory information to versioned_doc metadata
   // Example: `versioned_docs/version-1.1.6/projectA/readme.md` file with id `version-1.1.6-readme`
diff --git a/lib/version.js b/lib/version.js
index 0be72f6b4d..1cfa7221af 100755
--- a/lib/version.js
+++ b/lib/version.js
@@ -15,6 +15,7 @@ const chalk = require('chalk');
 const readMetadata = require('./server/readMetadata.js');
 const utils = require('./server/utils.js');
 const versionFallback = require('./server/versionFallback.js');
+const metadataUtils = require('./server/metadataUtils.js');
 const env = require('./server/env.js');
 
 const CWD = process.cwd();
@@ -85,7 +86,7 @@ files.forEach(file => {
     return;
   }
 
-  const res = readMetadata.extractMetadata(fs.readFileSync(file, 'utf8'));
+  const res = metadataUtils.extractMetadata(fs.readFileSync(file, 'utf8'));
   let metadata = res.metadata;
   // Don't version any docs without any metadata whatsoever.
   if (Object.keys(metadata).length === 0) {