diff --git a/v2/lib/load/blog.js b/v2/lib/load/blog.js
index 68680faf3199..1b95a7c8fe81 100644
--- a/v2/lib/load/blog.js
+++ b/v2/lib/load/blog.js
@@ -8,7 +8,7 @@
 const globby = require('globby');
 const path = require('path');
 const fs = require('fs-extra');
-const {parse, idx} = require('./utils');
+const {parse, idx, normalizeUrl} = require('./utils');
 
 function fileToUrl(fileName) {
   return fileName
@@ -47,7 +47,7 @@ async function loadBlog({blogDir, env, siteConfig}) {
       const fileString = await fs.readFile(source, 'utf-8');
       const {metadata: rawMetadata} = parse(fileString);
       const metadata = {
-        permalink: path.join(baseUrl, `blog`, fileToUrl(blogFileName)),
+        permalink: normalizeUrl([baseUrl, `blog`, fileToUrl(blogFileName)]),
         source,
         ...rawMetadata,
         date,
@@ -67,7 +67,10 @@ async function loadBlog({blogDir, env, siteConfig}) {
   /* eslint-disable */
   for (let page = 0; page < numberOfPage; page++) {
     blogMetadatas.push({
-      permalink: path.join(basePageUrl, `${page > 0 ? `page${page + 1}` : ''}`),
+      permalink: normalizeUrl([
+        basePageUrl,
+        `${page > 0 ? `page${page + 1}` : ''}`,
+      ]),
       language: defaultLangTag,
       isBlogPage: true,
       posts: blogMetadatas.slice(page * perPage, (page + 1) * perPage),
diff --git a/v2/lib/load/docs/metadata.js b/v2/lib/load/docs/metadata.js
index 477ed03e9c47..0885e3f9784d 100644
--- a/v2/lib/load/docs/metadata.js
+++ b/v2/lib/load/docs/metadata.js
@@ -7,7 +7,7 @@
 
 const fs = require('fs-extra');
 const path = require('path');
-const {getSubFolder, idx, parse} = require('../utils');
+const {getSubFolder, idx, parse, normalizeUrl} = require('../utils');
 
 function getLanguage(filepath, refDir, env) {
   const translationEnabled = idx(env, ['translation', 'enabled']);
@@ -135,9 +135,13 @@ module.exports = async function processMetadata(
         .replace(/:id/, metadata.id),
     );
   } else {
-    metadata.permalink = `${baseUrl}${docsUrl}/${langPart}${versionPart}${
-      metadata.id
-    }`;
+    metadata.permalink = normalizeUrl([
+      baseUrl,
+      docsUrl,
+      langPart,
+      versionPart,
+      metadata.id,
+    ]);
   }
 
   /* if version */
diff --git a/v2/lib/load/routes.js b/v2/lib/load/routes.js
index bd11232753fc..97b3013a6482 100644
--- a/v2/lib/load/routes.js
+++ b/v2/lib/load/routes.js
@@ -5,6 +5,8 @@
  * LICENSE file in the root directory of this source tree.
  */
 
+const {normalizeUrl} = require('./utils');
+
 async function genRoutesConfig({
   siteConfig = {},
   docsMetadatas = {},
@@ -33,7 +35,7 @@ async function genRoutesConfig({
       }`;
   }
 
-  const rootDocsUrl = baseUrl + docsUrl;
+  const rootDocsUrl = normalizeUrl([baseUrl, docsUrl]);
   const docsRoutes = `
   {
     path: '${rootDocsUrl}',
@@ -133,8 +135,9 @@ async function genRoutesConfig({
     `import BlogPage from '@theme/BlogPage';\n` +
     `import Pages from '@theme/Pages';\n` +
     `import NotFound from '@theme/NotFound';\n` +
-    `const routes = [${docsRoutes},
+    `const routes = [
       ${pagesMetadatas.map(genPagesRoute).join(',')},
+      ${docsRoutes},
       ${blogMetadatas.map(genBlogRoute).join(',')},
       ${notFoundRoute}\n];\n` +
     `export default routes;\n`
diff --git a/v2/lib/load/utils.js b/v2/lib/load/utils.js
index e6ab3677bd85..0fcc2ce5e2b1 100644
--- a/v2/lib/load/utils.js
+++ b/v2/lib/load/utils.js
@@ -77,6 +77,67 @@ function parse(fileString) {
   return {metadata, content};
 }
 
+function normalizeUrl(rawUrls) {
+  const urls = rawUrls;
+  const resultArray = [];
+
+  // If the first part is a plain protocol, we combine it with the next part.
+  if (urls[0].match(/^[^/:]+:\/*$/) && urls.length > 1) {
+    const first = urls.shift();
+    urls[0] = first + urls[0];
+  }
+
+  // There must be two or three slashes in the file protocol, two slashes in anything else.
+  if (urls[0].match(/^file:\/\/\//)) {
+    urls[0] = urls[0].replace(/^([^/:]+):\/*/, '$1:///');
+  } else {
+    urls[0] = urls[0].replace(/^([^/:]+):\/*/, '$1://');
+  }
+
+  // eslint-disable-next-line
+  for (let i = 0; i < urls.length; i++) {
+    let component = urls[i];
+
+    if (typeof component !== 'string') {
+      throw new TypeError(`Url must be a string. Received ${component}`);
+    }
+
+    if (component === '') {
+      // eslint-disable-next-line
+      continue;
+    }
+
+    if (i > 0) {
+      // Removing the starting slashes for each component but the first.
+      component = component.replace(/^[/]+/, '');
+    }
+    if (i < urls.length - 1) {
+      // Removing the ending slashes for each component but the last.
+      component = component.replace(/[/]+$/, '');
+    } else {
+      // For the last component we will combine multiple slashes to a single one.
+      component = component.replace(/[/]+$/, '/');
+    }
+
+    resultArray.push(component);
+  }
+
+  let str = resultArray.join('/');
+  // Each input component is now separated by a single slash except the possible first plain protocol part.
+
+  // remove trailing slash before parameters or hash
+  str = str.replace(/\/(\?|&|#[^!])/g, '$1');
+
+  // replace ? in parameters with &
+  const parts = str.split('?');
+  str = parts.shift() + (parts.length > 0 ? '?' : '') + parts.join('&');
+
+  // dedupe forward slashes
+  str = str.replace(/^\/+/, '/');
+
+  return str;
+}
+
 module.exports = {
   encodePath,
   generate,
@@ -84,5 +145,6 @@ module.exports = {
   fileToComponentName,
   getSubFolder,
   idx,
+  normalizeUrl,
   parse,
 };
diff --git a/v2/test/load/utils.test.js b/v2/test/load/utils.test.js
index 4488113fd039..5bc71269401e 100644
--- a/v2/test/load/utils.test.js
+++ b/v2/test/load/utils.test.js
@@ -11,6 +11,7 @@ import {
   fileToComponentName,
   idx,
   getSubFolder,
+  normalizeUrl,
 } from '@lib/load/utils';
 
 describe('load utils', () => {
@@ -104,4 +105,48 @@ describe('load utils', () => {
     expect(getSubFolder(testD, 'docs')).toBe('ro');
     expect(getSubFolder(testE, 'docs')).toBeNull();
   });
+
+  test('normalizeUrl', () => {
+    const asserts = [
+      {
+        input: ['/', '/'],
+        output: '/',
+      },
+      {
+        input: ['/', 'docs'],
+        output: '/docs',
+      },
+      {
+        input: ['/', 'docs', 'en', 'next', 'blog'],
+        output: '/docs/en/next/blog',
+      },
+      {
+        input: ['/test/', '/docs', 'ro', 'doc1'],
+        output: '/test/docs/ro/doc1',
+      },
+      {
+        input: ['', '/', 'ko', 'hello'],
+        output: '/ko/hello',
+      },
+      {
+        input: ['hello', 'world'],
+        output: 'hello/world',
+      },
+      {
+        input: ['http://www.google.com/', 'foo/bar', '?test=123'],
+        output: 'http://www.google.com/foo/bar?test=123',
+      },
+      {
+        input: ['http:', 'www.google.com///', 'foo/bar', '?test=123'],
+        output: 'http://www.google.com/foo/bar?test=123',
+      },
+      {
+        input: ['http://foobar.com', '', 'test'],
+        output: 'http://foobar.com/test',
+      },
+    ];
+    asserts.forEach(testCase => {
+      expect(normalizeUrl(testCase.input)).toBe(testCase.output);
+    });
+  });
 });
diff --git a/v2/website/siteConfig.js b/v2/website/siteConfig.js
index 8d4ad2538446..0fcc9fa65261 100644
--- a/v2/website/siteConfig.js
+++ b/v2/website/siteConfig.js
@@ -6,8 +6,8 @@
  */
 
 module.exports = {
-  title: 'docusaurus',
-  tagline: '📝⚡️ Transform your document (문서) to a website',
+  title: 'Docusaurus',
+  tagline: '⚡️ Painless static site generator',
   organizationName: 'facebook',
   projectName: 'docusaurus',
   baseUrl: '/',