diff --git a/assets/help_zh.html b/assets/help_zh.html
index b5ad7165..8533d2bc 100644
--- a/assets/help_zh.html
+++ b/assets/help_zh.html
@@ -6,18 +6,21 @@
 <a href="https://repo.or.cz/uweb.git/blob_plain/HEAD:/zh/index.html">repo</a>
 <a href="https://jamesfengcao.codeberg.page/zh/">codeberg</a>
 <a href="https://uwebzh.netlify.app/zh/">netlify</a>
+<a href="https://ik4ev-laaaa-aaaad-qd4sq-cai.ic.fleek.co/zh/">fleek</a>
 <a href="https://uweb.stormkit.dev/zh/">stormkit</a>
 <a href="https://uweb.surge.sh/zh/">surge</a>
 <a href="https://torappinfo.github.io/uweb/zh/">github</a>
-<a href="i:0m04:https://code.aliyun.com/torappinfo/uweb/raw/master/redirect.html">试验</a>
-<a href="i:0m04:https://gitcode.net/spoony1971/uweb/-/raw/master/redirect.html">试验2</a>
-<a href="i:0mi:5h:https://jihulab.com/torappinfo/uweb/-/raw/master/redirect.html">试验3</a>
-<a href="i:0m04:https://gitea.com/torappinfo/uweb/raw/branch/master/redirect.html">试验4</a>
+<a href="i:0mi:5h:https://gitee.com/jamesfengcao/uweb/raw/master/redirect.html">试验</a>
+<a href="i:0m04:https://gitcode.net/spoony1971/uweb/-/raw/master/redirect.html">试验1</a>
+<a href="i:0mi:5h:https://jihulab.com/torappinfo/uweb/-/raw/master/redirect.html">试验2</a>
+<a href="i:0m04:https://www.gitclone.com/gogs/jamesfengcao/uweb/raw/master/redirect.html">试验3</a>
+<a href="i:0m04:https://code.aliyun.com/torappinfo/uweb/raw/master/redirect.html">试验4</a>
+<a href="i:0m04:https://gitea.com/torappinfo/uweb/raw/branch/master/redirect.html">试验5</a>
+
+<a href="i:0mi:5h?format=raw::https://sourceforge.net/p/uwebbrowser/code/ci/master/tree/redirect.html">试验6</a>
+<a href="i:0mi:5h?format=raw::https://rocketgit.com/user/torappinfo/uweb/source/tree/branch/master/blob_download/redirect.html">试验7</a>
+<a href="i:0m04:https://notabug.org/torappinfo/uweb/raw/master/redirect.html">试验8</a>
 
-<a href="i:0mi:5h?format=raw::https://sourceforge.net/p/uwebbrowser/code/ci/master/tree/redirect.html">试验5</a>
-<a href="i:0mi:5h?format=raw::https://rocketgit.com/user/torappinfo/uweb/source/tree/branch/master/blob_download/redirect.html">试验6</a>
-<a href="i:0m04:https://notabug.org/torappinfo/uweb/raw/master/redirect.html">试验7</a>
-<a href="i:0mi:5h:https://gitee.com/jamesfengcao/uweb/raw/master/redirect.html">试验8</a>
 <a href="i:0mi:5h:https://git.launchpad.net/uweb/plain/redirect.html">试验9</a>
 <a href="i:0m04:https://framagit.org/torappinfo/uweb/-/raw/master/redirect.html">试验10</a>
 <a href="i:0m04:https://agit.ai/jamesfengcao/uweb/raw/branch/master/redirect.html">试验11</a>
@@ -28,6 +31,7 @@
 <a href="i:0mi:5h:https://pagure.io/uweb/raw/master/f/redirect.html">试验16</a>
 
 <a href="i:0mi:5h:https://fastly.jsdelivr.net/gh/torappinfo/uweb/redirect.html">试验17</a>
+
 <a href="https://torappinfo.bitbucket.io/zh/">bitbucket</a>
 <a href="https://uweb-zh.vercel.app/zh/">vercel</a>
 <a href="https://uwebzh.pages.dev/zh/">pages</a>
diff --git a/en/index.html b/en/index.html
index 89870101..42725482 100644
--- a/en/index.html
+++ b/en/index.html
@@ -1,7 +1,7 @@
 <!doctype html>
 <html>
   <head>
-	<meta name="generator" content="Hugo 0.104.2" />
+	<meta name="generator" content="Hugo 0.104.3" />
     <meta charset="utf-8">
     <meta name="viewport" content="width=device-width, initial-scale=1" />
     
diff --git a/en/links/index.html b/en/links/index.html
index d86d9477..70d44034 100644
--- a/en/links/index.html
+++ b/en/links/index.html
@@ -148,6 +148,10 @@ Wait to receive url from the network and open the url (not showing ip)</p>
 Launch app for specific <action>.</p>
 </li>
 <li>
+<p>i:1e<br>
+Open file manager for removable SDCARD download folder</p>
+</li>
+<li>
 <p>i:50[path]:[string]<br>
 Save [string] as file</p>
 </li>
diff --git a/en/searchurl/mdict/mdict-common.js b/en/searchurl/mdict/mdict-common.js
new file mode 100644
index 00000000..20f11a63
--- /dev/null
+++ b/en/searchurl/mdict/mdict-common.js
@@ -0,0 +1,22 @@
+var MCommon = (function () {
+  return {
+    /**
+     * Get file extension.
+     */
+    getExtension: function (filename, defaultExt) {
+      return /(?:\.([^.]+))?$/.exec(filename)[1] || defaultExt;
+    },
+    
+     /**
+      * Regular expression to strip key if dictionary's "StripKey" attribute is true. 
+      */
+    REGEXP_STRIPKEY: {
+      'mdx' : /[()., '/\\@_-]()/g,
+      'mdd' : /([.][^.]*$)|[()., '/\\@_-]/g        // strip '.' before file extension that is keeping the last period
+    },
+
+    log: function() {
+      console.log.apply(console, [].slice.apply(arguments));
+    }
+  };
+}());
diff --git a/en/searchurl/mdict/mdict-parser.js b/en/searchurl/mdict/mdict-parser.js
new file mode 100644
index 00000000..658fe218
--- /dev/null
+++ b/en/searchurl/mdict/mdict-parser.js
@@ -0,0 +1,938 @@
+//define(['pako', 'lzo', 'ripemd128', 'bluebird', 'mdict-parseXml', 'mdict-MCommon']
+//pako, lzo, ripemd128, Promise, parseXml, MCommon
+var parseXml = function (str) {
+  return (new DOMParser()).parseFromString(str, 'text/xml');
+};
+var MParser = (function (){
+  // Value of undefined.
+  var UNDEFINED = void 0;
+  
+  // A shared UTF-16LE text decorder used to read dictionary header string.
+  var UTF_16LE = new TextDecoder('utf-16le');
+  
+  /**
+   * Return the first argument as result.
+   * This function is used to simulate consequence, i.e. read data and return it, then forward to a new position.
+   * @param any data or function call
+   * @return the first arugment
+   */
+  function conseq(/* args... */) { return arguments[0]; }
+
+  /*
+   * Decrypt encrypted data block of keyword index (attrs.Encrypted = "2").
+   * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#keyword-index-encryption
+   * @param buf an ArrayBuffer containing source data
+   * @param key an ArrayBuffer holding decryption key, which will be supplied to ripemd128() before decryption
+   * @return an ArrayBuffer carrying decrypted data, occupying the same memory space of source buffer
+   */
+  function decrypt(buf, key) {
+    key = ripemd128(key);
+    var byte, keylen = key.length, prev = 0x36, i = 0, len = buf.length;
+    for (; i < len; i++) {
+      byte = buf[i];
+      byte = ((byte >> 4) | (byte << 4) );                  // & 0xFF;  <-- it's already a byte
+      byte = byte  ^ prev ^ (i & 0xFF) ^ key[i % keylen];
+      prev = buf[i];
+      buf[i] = byte;
+    }
+    return buf;
+  }
+  
+  /**
+   * For sliceThen(..).exec(proc, ..), mark what proc function returns is multiple values 
+   * to be passed to further Promise#spread(..) call.
+   */
+  function spreadus() {
+    var args = Array.prototype.slice.apply(arguments);
+    args._spreadus_ = true;
+    return args;
+  }
+  
+  /**
+   * Slice part of a file/blob object, return a promise object which will resolve to an ArrayBuffer to feed subsequent process.
+   * The returned promise object is extened with an exec(proc, args...) method which can be chained with further process.
+   * @param file file or blob object
+   * @param offset start position to slice
+   * @param len length to slice
+   * @return a promise object which will resolve to an ArrayBuffer containing data been read
+   */
+  function sliceThen(file, offset, len) {
+    var p = new Promise(function(_resolve) {
+      var reader = new FileReader();
+      reader.onload = function() { _resolve(reader.result); }
+      reader.readAsArrayBuffer(file.slice(offset, offset + len));
+    });
+
+    /**
+     * Call proc with specified arguments prepending with sliced file/blob data (ArrayBuffer) been read.
+     * @param the first argument is a function to be executed
+     * @param other optional arguments are passed to the function following auto supplied input ArrayBuffer
+     * @return a promise object which can be chained with further process through spread() method
+     */
+    p.exec = function(proc /*, args... */) {
+      var args = Array.prototype.slice.call(arguments, 1);
+      return p.then(function(data) {
+          args.unshift(data);
+          var ret = proc.apply(null, args);
+          return resolve(ret !== UNDEFINED && ret._spreadus_ ? ret : [ret]);
+      });
+    };
+        
+    return p;
+  }
+  
+  /**
+   * Wrap value as a resolved promise.
+   */
+  function resolve(value) { return Promise.resolve(value); }
+  
+  /**
+   * Wrap value as a rejected promise.
+   */
+  function reject(reason) { return Promise.reject(reason); }
+  
+  /**
+   * Harvest any resolved promises, if all failed then return reasons. 
+   */
+  function harvest(outcomes) {
+    return Promise.settle(outcomes).then(function(results) {
+      if (results.length === 0) {
+        return reject("** NOT FOUND **");
+      }
+      
+      var solved = [], failed = [];
+      for (var i = 0; i < results.length; i++) {
+        if (results[i].isResolved()) {
+          solved.push(results[i].value());
+        } else {
+          failed.push(results[i].reason());
+        }
+      }
+      return solved.length ? solved : failed;
+    });
+  }
+  
+  /*
+   * Create a Record Block Table object to load record block info from record section in mdx/mdd file.
+   * Retrived data is stored in an Uint32Array which contains N pairs of (offset_comp, offset_decomp) value,
+   * where N is number of record blocks.
+   *
+   * When looking up a given key for its definition:
+   *   1. Search KEY_INDEX to locate keyword block containing the given key.
+   *   2. Scanning the found keyword block to get its record offset and size.
+   *   3. Search RECORD_BLOCK_TABLE to get record block containing the record.
+   *   4. Load the found record block, using its offset and size to retrieve record content.
+   *
+   * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#record-section
+   */
+  function createRecordBlockTable() {
+    var pos = 0, // current position
+        arr;     // backed Uint32Array
+    return {
+      // Allocate required ArrayBuffer for storing record block table, where len is number of record blocks.
+      alloc:  function(len) { 
+                arr = new Uint32Array(len * 2);
+              },
+      // Store offset pair value (compressed & decompressed) for a record block
+      // NOTE: offset_comp is absolute offset counted from start of mdx/mdd file.
+      put:    function(offset_comp, offset_decomp) { 
+                arr[pos++] = offset_comp; arr[pos++] = offset_decomp;
+              },
+      // Given offset of a keyword after decompression, return a record block info containing it, else undefined if not found.
+      find:   function(keyAt) {
+                var hi = (arr.length >> 1) - 1, lo = 0, i = (lo + hi) >> 1, val = arr[(i << 1) + 1];
+
+                if (keyAt > arr[(hi << 1) + 1] || keyAt < 0) {
+                  return;
+                }
+
+                while (true) {                    
+                  if (hi - lo <= 1) {
+                    if (i < hi) {
+                      return {
+                        block_no:     i,
+                        comp_offset:  arr[i <<= 1],
+                        comp_size:    arr[i + 2] - arr[i],
+                        decomp_offset:arr[i + 1],
+                        decomp_size:  arr[i + 3] - arr[i + 1]
+                      };
+                    } else {
+                      return;
+                    }
+                  }
+
+                  (keyAt < val)  ? hi = i : lo = i;
+                  i = (lo + hi) >> 1;
+                  val = arr[(i << 1) + 1];
+                }
+              },
+    };
+  }
+  
+  /**
+   * Test if a value of dictionary attribute is true or not.
+   */
+  function isTrue(v) {
+    v = ((v || false) + '').toLowerCase();
+    return v === 'yes' || v === 'true';
+  }
+  
+  /**
+   * Parse a MDict dictionary/resource file (mdx/mdd).
+   * @param file a File/Blob object
+   * @param ext file extension, mdx/mdd
+   * @return a Promise object which will resolve to a lookup function.
+   */
+  function parse_mdict(file, ext) {
+
+    var KEY_INDEX,                                       // keyword index array
+        RECORD_BLOCK_TABLE = createRecordBlockTable();   // record block table
+
+    var attrs = {},     // storing dictionary attributes
+        _v2,            // true if enginge version > 2
+        _bpu,           // bytes per unit when converting text size to byte length for text data
+        _tail,          // need to skip extra tail bytes after decoding text
+        _decoder,       // text decorder
+
+        _decryptors  = [false, false],
+                        // [keyword_header_decryptor, keyword_index_decryptor], only keyword_index_decryptor is supported
+        
+        _searchTextLen, // search NUL to get text length
+        
+        _readShort   = function(scanner) { return scanner.readUint8(); },
+                        // read a "short" number representing kewword text size, 8-bit for version < 2, 16-bit for version >= 2
+
+        _readNum     = function(scanner) { return scanner.readInt(); },
+                        // Read a number representing offset or data block size, 16-bit for version < 2, 32-bit for version >= 2
+        
+        _checksum_v2 = function() {},
+                        // Version >= 2.0 only checksum
+        
+        _adaptKey    = function(key) { return key; },
+                        // adapt key by converting to lower case or stripping punctuations according to dictionary attributes (KeyCaseSensitive, StripKey)
+        
+        _slice       = sliceThen.bind(null, file);
+                        // bind sliceThen() with file argument          
+        
+    /**
+     * Config scanner according to dictionary attributes.
+     */
+    function config() {
+      attrs.Encoding = attrs.Encoding || 'UTF-16';
+      
+      _searchTextLen = (attrs.Encoding === 'UTF-16') 
+                          ?   function(dv, offset) {
+                                offset = offset;
+                                var mark = offset;
+                                while (dv.getUint16(offset)) { offset+= _bpu /* scan for \u0000 */ };
+                                return offset - mark;
+                          } : function(dv, offset) {
+                                offset = offset;
+                                var mark = offset;
+                                while (dv.getUint8(offset++)) { /* scan for NUL */ }
+                                return offset - mark - 1;
+                          };
+      
+      _decoder = new TextDecoder(attrs.Encoding || 'UTF-16LE');
+
+      _bpu = (attrs.Encoding === 'UTF-16') ? 2 : 1;
+      
+      if (parseInt(attrs.GeneratedByEngineVersion, 10) >= 2.0) {
+        _v2 = true;
+        _tail = _bpu;
+
+          // HUGE dictionary file (>4G) is not supported, take only lower 32-bit
+        _readNum     = function(scanner) { return scanner.forward(4), scanner.readInt(); };
+        _readShort   = function(scanner) { return scanner.readUint16(); };
+        _checksum_v2 = function(scanner) { return scanner.checksum(); };
+      } else {
+        _tail = 0;
+      }
+      
+      // keyword index decrypted?
+      if (attrs.Encrypted & 0x02) {
+        _decryptors[1] = decrypt; 
+      }
+      
+      var regexp = MCommon.REGEXP_STRIPKEY[ext];
+      if (isTrue(attrs.KeyCaseSensitive)) {
+        _adaptKey = isTrue(attrs.StripKey) 
+                      ? function(key) { return key.replace(regexp, '$1'); }
+                      : function(key) { return key; };
+      } else {
+        _adaptKey = isTrue(attrs.StripKey || (_v2 ? '' : 'yes')) 
+                      ? function(key) { return key.toLowerCase().replace(regexp, '$1'); }
+                      : function(key) { return key.toLowerCase(); };
+      }
+    }
+    
+    // Read data in current offset from target data ArrayBuffer
+    function Scanner(buf, len) {
+      var offset = 0, dv = new DataView(buf);
+
+      var methods = {
+        // target data size in bytes
+        size:     function() { return len || buf.byteLength; },
+        // update offset to new position
+        forward:  function(len) { return offset += len; },
+        // return current offset
+        offset:   function() { return offset; },
+        
+        // MDict file format uses big endian to store number
+        
+        // 32-bit unsigned int
+        readInt:    function() { return conseq(dv.getUint32(offset, false), this.forward(4)); },
+        readUint16: function() { return conseq(dv.getUint16(offset, false), this.forward(2)); },
+        readUint8:  function() { return conseq(dv.getUint8(offset, false),  this.forward(1)); },
+        
+        // Read a "short" number representing keyword text size, 8-bit for version < 2, 16-bit for version >= 2
+        readShort:  function() { return _readShort(this); },
+        // Read a number representing offset or data block size, 16-bit for version < 2, 32-bit for version >= 2
+        readNum:    function() { return _readNum(this); },
+
+        readUTF16:  function(len) { return conseq(UTF_16LE.decode(new Uint8Array(buf, offset, len)), this.forward(len)); },
+        
+        // Read data to an Uint8Array and decode it to text with specified encoding.
+        // Text length in bytes is determined by searching terminated NUL.
+        // NOTE: After decoding the text, it is need to forward extra "tail" bytes according to specified encoding. 
+        readText: function() {
+          var len = _searchTextLen(dv, offset);
+          return conseq(_decoder.decode(new Uint8Array(buf, offset, len)), this.forward(len + _bpu));
+        },
+        // Read data to an Uint8Array and decode it to text with specified encoding.
+        // @param len length in basic unit, need to multiply byte per unit to get length in bytes
+        // NOTE: After decoding the text, it is need to forward extra "tail" bytes according to specified encoding. 
+        readTextSized: function(len) {
+          len *= _bpu;
+          return conseq(_decoder.decode(new Uint8Array(buf, offset, len)), this.forward(len + _tail));
+        },
+        
+        // Skip checksum, just ignore it anyway.
+        checksum:    function() { this.forward(4); },
+        // Version >= 2.0 only
+        checksum_v2: function() { return _checksum_v2(this); },
+
+        // Read data block of keyword index, key block or record content.
+        // These data block are maybe in compressed (gzip or lzo) format, while keyword index maybe be encrypted.
+        // @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#compression (with typo mistake)
+        readBlock: function(len, expectedBufSize, decryptor) {
+          var comp_type = dv.getUint8(offset, false);  // compression type, 0 = non, 1 = lzo, 2 = gzip
+          if (comp_type === 0) {
+            if (_v2) {
+              this.forward(8);  // for version >= 2, skip comp_type (4 bytes with tailing \x00) and checksum (4 bytes)
+            }
+            return this;
+          } else {
+            // skip comp_type (4 bytes with tailing \x00) and checksum (4 bytes)
+            offset += 8; len -= 8;
+            var tmp = new Uint8Array(buf, offset, len);
+            if (decryptor) {
+              var passkey = new Uint8Array(8);
+              passkey.set(new Uint8Array(buf, offset - 4, 4));  // key part 1: checksum
+              passkey.set([0x95, 0x36, 0x00, 0x00], 4);         // key part 2: fixed data
+              tmp = decryptor(tmp, passkey);
+            }
+            
+            tmp = comp_type === 2 ? pako.inflate(tmp) : lzo.decompress(tmp, expectedBufSize, 1308672);
+            this.forward(len);
+            return Scanner(tmp.buffer, tmp.length);
+          }
+        },
+        
+        // Read raw data as Uint8Array from current offset with specified length in bytes
+        readRaw: function(len) {
+          return conseq(new Uint8Array(buf, offset, len), this.forward(len === UNDEFINED ? buf.length - offset : len));
+        },
+      };
+
+      return Object.create(methods);
+    }
+    
+    /**
+     * Read the first 4 bytes of mdx/mdd file to get length of header_str.
+     * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#file-structure
+     * @param input sliced file (start = 0, length = 4)
+     * @return length of header_str
+     */
+    function read_file_head(input) {
+      return Scanner(input).readInt();
+    }
+
+    /**
+     * Read header section, parse dictionary attributes and config scanner according to engine version attribute.
+     * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#header-section
+     * @param input sliced file (start = 4, length = len + 48), header string + header section (max length 48)
+     * @param len lenghth of header_str
+     * @return [remained length of header section (header_str and checksum, = len + 4), original input]
+     */
+    function read_header_sect(input, len) {
+      var scanner = Scanner(input),
+          header_str = scanner.readUTF16(len).replace(/\0$/, ''); // need to remove tailing NUL
+
+      // parse dictionary attributes
+      var xml = parseXml(header_str).querySelector('Dictionary, Library_Data').attributes;
+
+      for (var i = 0, item; i < xml.length; i++) {
+        item = xml.item(i);
+        attrs[item.nodeName] = item.nodeValue;
+      }
+
+      attrs.Encrypted = parseInt(attrs.Encrypted, 10) || 0;
+
+      MCommon.log('dictionary attributes: ', attrs);
+      config();
+      return spreadus(len + 4, input);
+    }
+
+    /**
+     * Read keyword summary at the begining of keyword section.
+     * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#keyword-section
+     * @param input sliced file, same as input passed to read_header_sect()
+     * @param offset start position of keyword section in sliced file, equals to length of header string plus checksum.\
+     * @return keyword_sect object
+     */
+    function read_keyword_summary(input, offset) {
+      var scanner = Scanner(input);
+      scanner.forward(offset);
+      return {
+        num_blocks:           scanner.readNum(),
+        num_entries:          scanner.readNum(),
+        key_index_decomp_len: _v2 && scanner.readNum(),  // Ver >= 2.0 only
+        key_index_comp_len:   scanner.readNum(),
+        key_blocks_len:       scanner.readNum(),
+        chksum:               scanner.checksum_v2(),
+        // extra field
+        len:                  scanner.offset() - offset,  // actual length of keyword section, varying with engine version attribute
+      };
+    }
+
+    /**
+     * Read keyword index part of keyword section. 
+     * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#keyword-header-encryption
+     * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#keyword-index
+     * @param input sliced file, remained part of keyword section after keyword summary which can also be used to read following key blocks.
+     * @param keyword_summary 
+     * @return [keyword_summary, array of keyword index]
+     */
+    function read_keyword_index(input, keyword_summary) {
+      var scanner = Scanner(input).readBlock(keyword_summary.key_index_comp_len, keyword_summary.key_index_decomp_len, _decryptors[1]),
+          keyword_index = Array(keyword_summary.num_blocks),
+          offset = 0;
+      
+      for (var i = 0, size; i < keyword_summary.num_blocks; i++) {
+        keyword_index[i] = {
+          num_entries: conseq(scanner.readNum(), size = scanner.readShort()),
+// UNUSED, can be ignored        
+//          first_size:  size = scanner.readShort(),
+          first_word:  conseq(scanner.readTextSized(size), size = scanner.readShort()),
+// UNUSED, can be ignored
+//          last_size:   size = scanner.readShort(),
+          last_word:   scanner.readTextSized(size),
+          comp_size:   size = scanner.readNum(),
+          decomp_size: scanner.readNum(),
+          // extra fields
+          offset: offset,     // offset of the first byte for the target key block in mdx/mdd file
+          index: i            // index of this key index, used to search previous/next block
+        };
+        offset += size;
+      }
+      return spreadus(keyword_summary, keyword_index);
+    }
+
+    /**
+     * Read keyword entries inside a keyword block and fill KEY_TABLE.
+     * @param scanner scanner object to read key entries, which starts at begining of target key block
+     * @param kdx corresponding keyword index object
+     * NOTE: no need to read keyword block anymore, for debug only.
+     */
+    function read_key_block(scanner, kdx) {
+      var scanner = scanner.readBlock(kdx.comp_size, kdx.decomp_size);
+      for (var i = 0; i < kdx.num_entries; i++) {
+//        scanner.readNum(); scanner.readText();
+        var kk = [scanner.readNum(), scanner.readText()];
+//        console.log(scanner.readNum(), scanner.readText());
+      }
+    }
+    
+    /**
+     * Delay to scan key table, for debug onyl.
+     * @param slicedKeyBlock a promise object which will resolve to an ArrayBuffer containing keyword blocks 
+     *                       sliced from mdx/mdd file.
+     * @param num_entries number of keyword entries
+     * @param keyword_index array of keyword index
+     * @param delay time to delay for scanning key table
+     */
+    function willScanKeyTable(slicedKeyBlock, num_entries, keyword_index, delay) {
+      slicedKeyBlock.delay(delay).then(function (input) {
+        MCommon.log('scan key table...');
+        var scanner = Scanner(input);
+        for (var i = 0, size = keyword_index.length; i < size; i++) {
+          read_key_block(scanner, keyword_index[i]);
+        }
+
+        MCommon.log('KEY_TABLE loaded.');
+      });
+    }
+
+    /**
+     * Read record summary at the begining of record section.
+     * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#record-section
+     * @param input sliced file, start = begining of record section, length = 32 (max length of record summary)
+     * @param pos begining of record section
+     * @returj record summary object
+     */
+    function read_record_summary(input, pos) {
+      var scanner = Scanner(input),
+          record_summary = {
+            num_blocks:   scanner.readNum(),
+            num_entries:  scanner.readNum(),
+            index_len:    scanner.readNum(),
+            blocks_len:   scanner.readNum(),
+            // extra field
+            len:          scanner.offset(),   // actual length of record section (excluding record block index), varying with engine version attribute
+          };
+      
+      // start position of record block from head of mdx/mdd file
+      record_summary.block_pos = pos + record_summary.index_len + record_summary.len;
+
+      return record_summary;
+    }
+
+    /**
+     * Read record block index part in record section, and fill RECORD_BLOCK_TABLE
+     * @see https://github.com/zhansliu/writemdict/blob/master/fileformat.md#record-section
+     * @param input sliced file, start = begining of record block index, length = record_summary.index_len
+     * @param record_summary record summary object
+     */
+    function read_record_block(input, record_summary) {
+      var scanner = Scanner(input),
+          size = record_summary.num_blocks,
+          record_index = Array(size),
+          p0 = record_summary.block_pos, 
+          p1 = 0;
+
+      RECORD_BLOCK_TABLE.alloc(size + 1);
+      for (var i = 0, rdx; i < size; i++) {
+        record_index[i] = rdx = {
+          comp_size:   scanner.readNum(),
+          decomp_size: scanner.readNum()
+        };
+        RECORD_BLOCK_TABLE.put(p0, p1);
+        p0 += rdx.comp_size;
+        p1 += rdx.decomp_size;
+      }
+      RECORD_BLOCK_TABLE.put(p0, p1);
+    }
+    
+    /**
+     * Read definition in text for given keyinfo object.
+     * @param input record block sliced from the file
+     * @param block record block index 
+     * @param keyinfo a object with property of record's offset and optional size for the given keyword
+     * @return definition in text
+     */
+    function read_definition(input, block, keyinfo) {
+      var scanner = Scanner(input).readBlock(block.comp_size, block.decomp_size);
+      scanner.forward(keyinfo.offset - block.decomp_offset);
+      return scanner.readText();
+    }
+    
+    /**
+     * Following link to find actual definition of keyword.
+     * @param definition maybe starts with "@@@LINK=" which links to another keyword 
+     * @param lookup search function
+     * @return resolved actual definition
+     */
+    function followLink(definition, lookup) {
+      return (definition.substring(0, 8) !== '@@@LINK=') 
+                ? definition
+                : lookup(definition.substring(8));
+    }
+
+    /**
+     * Read content in ArrayBuffer for give keyinfo object
+     * @param input record block sliced from the file
+     * @param block record block index 
+     * @param keyinfo a object with property of record's offset and optional size for the given keyword
+     * @return an ArrayBuffer containing resource of image/audio/css/font etc.
+     */
+    function read_object(input, block, keyinfo) {
+      if (input.byteLength > 0) {
+        var scanner = Scanner(input).readBlock(block.comp_size, block.decomp_size);
+        scanner.forward(keyinfo.offset - block.decomp_offset);
+        return scanner.readRaw(keyinfo.size);
+      } else {
+        throw '* OUT OF FILE RANGE * ' + keyinfo + ' @offset=' + block.comp_offset;
+      }
+    }
+    
+    /**
+     * Find word definition for given keyinfo object.
+     * @param keyinfo a object with property of record's offset and optional size for the given keyword
+     * @return a promise object which will resolve to definition in text. Link to other keyword is followed to get actual definition.
+     */
+    function findWord(keyinfo) {
+      var block = RECORD_BLOCK_TABLE.find(keyinfo.offset);
+      return _slice(block.comp_offset, block.comp_size)
+                .exec(read_definition, block, keyinfo)
+                  .spread(function (definition) { return resolve(followLink(definition, LOOKUP.mdx)); });
+    }
+    
+    /**
+     * Find resource (image, sound etc.) for given keyinfo object.
+     * @param keyinfo a object with property of record's offset and optional size for the given keyword
+     * @return a promise object which will resolve to an ArrayBuffer containing resource of image/audio/css/font etc.
+     * TODO: Follow link, maybe it's too expensive and a rarely used feature?
+     */
+    function findResource(keyinfo) {
+      var block = RECORD_BLOCK_TABLE.find(keyinfo.offset);
+      return _slice(block.comp_offset, block.comp_size)
+                .exec(read_object, block, keyinfo)
+                  .spread(function (blob) { return resolve(blob); });
+    }
+    
+    //------------------------------------------------------------------------------------------------
+    // Implementation for look-up
+    //------------------------------------------------------------------------------------------------
+    var slicedKeyBlock,
+        _cached_keys,             // cache latest keys 
+        _trail,                   // store latest visited record block & position when search for candidate keys
+        mutual_ticket = 0;        // a oneway increased ticket used to cancel unfinished pattern match
+
+    
+    /**
+      * Reduce the key index array to an element which contains or is the nearest one matching a given phrase.
+      */
+    function reduce(arr, phrase) {
+      var len = arr.length;
+      if (len > 1) {
+        len = len >> 1;
+        return phrase > _adaptKey(arr[len - 1].last_word)
+                  ? reduce(arr.slice(len), phrase)
+                  : reduce(arr.slice(0, len), phrase);
+      } else {
+        return arr[0];
+      }
+    }
+    
+    /**
+      * Reduce the array to index of an element which contains or is the nearest one matching a given phrase.
+      */
+    function shrink(arr, phrase) {
+      var len = arr.length, sub;
+      if (len > 1) {
+        len = len >> 1;
+        var key = _adaptKey(arr[len]);
+        if (phrase < key) {
+          sub = arr.slice(0, len);
+          sub.pos = arr.pos;
+        } else {
+          sub = arr.slice(len);
+          sub.pos = (arr.pos || 0) + len;
+        }
+        return shrink(sub, phrase);
+      } else {
+        return (arr.pos || 0) + (phrase <= _adaptKey(arr[0]) ? 0 : 1);
+      }
+    }
+
+    /**
+     * Load keys for a keyword index object from mdx/mdd file.
+     * @param kdx keyword index object
+     */
+    function loadKeys(kdx) {
+      if (_cached_keys && _cached_keys.pilot === kdx.first_word) {
+        return resolve(_cached_keys.list);
+      } else {
+        return slicedKeyBlock.then(function(input) {
+          var scanner = Scanner(input), list = Array(kdx.num_entries);
+          scanner.forward(kdx.offset);
+          scanner = scanner.readBlock(kdx.comp_size, kdx.decomp_size);
+              
+          for (var i = 0; i < kdx.num_entries; i++) {
+            var offset = scanner.readNum();
+            list[i] = new Object(scanner.readText());
+            list[i].offset = offset;
+            if (i > 0) {
+              list[i - 1].size = offset - list[i - 1].offset;
+            }
+          }
+          _cached_keys = {list: list, pilot: kdx.first_word};
+          return list;
+        });
+      }
+    }
+        
+    /**
+     * Search for the first keyword match given phrase.
+     */
+    function seekVanguard(phrase) {
+      phrase = _adaptKey(phrase);
+      var kdx = reduce(KEY_INDEX, phrase);
+
+      // look back for the first record block containing keyword for the specified phrase
+      if (phrase <= _adaptKey(kdx.last_word)) {
+        var index = kdx.index - 1, prev;
+        while (prev = KEY_INDEX[index]) {
+          if (_adaptKey(prev.last_word) !== _adaptKey(kdx.last_word)) {
+            break;
+          }
+          kdx = prev;
+          index--;
+        }
+      }
+
+      return loadKeys(kdx).then(function (list) {
+          var idx = shrink(list, phrase);
+          // look back for the first matched keyword position
+          while (idx > 0) {
+            if (_adaptKey(list[--idx]) !== _adaptKey(phrase)) {
+              idx++;
+              break;
+            }
+          }        
+          return [kdx, Math.min(idx, list.length - 1), list];
+        });
+    }
+    
+    // TODO: have to restrict max count to improve response
+    /**
+     * Append more to word list according to a filter or expected size.
+     */
+    function appendMore(word, list, nextKdx, expectedSize, filter, ticket) {
+      if (ticket !== mutual_ticket) {
+        throw 'force terminated';
+      }
+
+      if (filter) {
+        if (_trail.count < expectedSize && nextKdx && nextKdx.first_word.substr(0, word.length) === word) {
+          return loadKeys(nextKdx).delay(30).then(function(more) {
+            MCommon.log(nextKdx);
+            _trail.offset = 0;
+            _trail.block = nextKdx.index;
+            Array.prototype.push.apply(list, more.filter(filter, _trail));
+            return appendMore(word, list, KEY_INDEX[nextKdx.index + 1], expectedSize, filter, ticket);
+          });
+        } else {
+          if (list.length === 0) {
+            _trail.exhausted = true;
+          }
+          return resolve(list);
+        }
+      } else {
+        var shortage = expectedSize - list.length;
+        if (shortage > 0 && nextKdx) {
+console.log('go next', nextKdx);
+          _trail.block = nextKdx.index;
+          return loadKeys(nextKdx).then(function(more) {
+            _trail.offset = 0;
+            _trail.pos = Math.min(shortage, more.length);
+            Array.prototype.push.apply(list, more.slice(0, shortage));
+console.log('$$ ' + more[shortage - 1], shortage);                
+            return appendMore(word, list, KEY_INDEX[nextKdx.index + 1], expectedSize, filter, ticket);
+          });
+        } else {
+          if (_trail.pos > expectedSize) {
+            _trail.pos = expectedSize;
+          }
+          list = list.slice(0, expectedSize);
+          _trail.count = list.length;
+          _trail.total += _trail.count;
+          return resolve(list);
+        }
+      }
+    }
+        
+    function followUp() {
+      var kdx = KEY_INDEX[_trail.block];
+      return loadKeys(kdx).then(function (list) {
+                return [kdx, Math.min(_trail.offset + _trail.pos, list.length - 1), list];
+              });
+    }
+      
+    function matchKeys(phrase, expectedSize, follow) {
+      expectedSize = Math.max(expectedSize || 0, 10);
+      var str = phrase.trim().toLowerCase(),
+          m = /([^?*]+)[?*]+/.exec(str), 
+          word;
+      if (m) {
+        word = m[1];
+        var wildcard = new RegExp('^' + str.replace(/([\.\\\+\[\^\]\$\(\)])/g, '\\$1').replace(/\*+/g, '.*').replace(/\?/g, '.') + '$'),
+            tester = phrase[phrase.length - 1] === ' ' 
+                        ? function(s) { return wildcard.test(s); } 
+                        : function(s) { return wildcard.test(s) && !/ /.test(s); },
+            filter = function (s, i) { 
+                       if (_trail.count < expectedSize && tester(s)) {
+                         _trail.count++;
+                         _trail.total++;
+                         _trail.pos = i + 1;
+                         return true;
+                       }
+                       return false;
+                     };
+      } else {
+        word = phrase.trim();
+      }
+      
+      if (_trail && _trail.phrase !== phrase) {
+        follow = false;
+      }
+      
+      if (follow && _trail && _trail.exhausted) {
+        return resolve([]);
+      }
+      
+      var startFrom = follow && _trail ? followUp() : seekVanguard(word);
+      
+      return startFrom.spread(function(kdx, idx, list) {
+console.log('start  ', kdx);
+        list = list.slice(idx);
+        _trail = {phrase: phrase, 
+                  block:  kdx.index, 
+                  offset: idx, 
+                  pos:    list.length, 
+                  count:  0,
+                  total:  follow ? _trail && _trail.total || 0 : 0
+                };
+        if (filter) {
+          list = list.filter(filter, _trail);
+        }
+        return appendMore(word, list, KEY_INDEX[kdx.index + 1], expectedSize, filter, ++mutual_ticket)
+                  .then(function(result) {
+                    if (_trail.block === KEY_INDEX.length - 1) {
+                      if (_trail.offset + _trail.pos >= KEY_INDEX[_trail.block].num_entries) {
+                        _trail.exhausted = true;
+console.log('EXHAUSTED!!!!');
+                      }
+                    }
+console.log('trail: ', _trail);          
+                    return result;
+                  });
+      });
+    };
+    
+    
+    /**
+     * Match the first element in list with given offset.
+     */
+    function matchOffset(list, offset) {
+      return list.some(function(el) { return el.offset === offset ? list = [el] : false; }) ? list : [];
+    }
+    
+    // Lookup functions
+    var LOOKUP = {
+      /**
+       * @param query 
+       *          String
+       *          {phrase: .., max: .., follow: true} object
+       */
+      mdx: function(query) {
+        if (typeof query === 'string' || query instanceof String) {
+          _trail = null;
+          var word = query.trim().toLowerCase(), offset = query.offset;
+
+          return seekVanguard(word).spread(function(kdx, idx, list) {
+            list = list.slice(idx);
+            if (offset !== UNDEFINED) {
+              list = matchOffset(list, offset);
+            } else {
+              list = list.filter(function(el) { return el.toLowerCase() === word; });
+            }
+            return harvest(list.map(findWord));
+          });
+        } else {
+          return matchKeys(query.phrase, query.max, query.follow);
+        }
+      },
+      
+      // TODO: chain multiple mdd file
+      mdd: function(phrase) {
+        var word = phrase.trim().toLowerCase();
+        word = '\\' + word.replace(/(^[/\\])|([/]$)/, '');
+        word = word.replace(/\//g, '\\');
+        return seekVanguard(word).spread(function(kdx, idx, list) {
+          return list.slice(idx).filter(function(one) {
+            return one.toLowerCase() === word;
+          });
+        }).then(function(candidates) {
+          if (candidates.length === 0) {
+            throw '*RESOURCE NOT FOUND* ' + phrase;
+          } else {
+            return findResource(candidates[0]);
+          }
+        });
+      }
+    };    
+    
+    // ------------------------------------------
+    // start to load mdx/mdd file
+    // ------------------------------------------
+    MCommon.log('start to load ' + file.name);
+    
+    var pos = 0;
+
+    // read first 4 bytes to get header length
+    return _slice(pos, 4).exec(read_file_head).spread(function(len) {
+      pos += 4;                                   // start of header string in header section
+      return _slice(pos, len + 48)
+                .exec(read_header_sect, len);
+
+    }).spread(function(header_remain_len, input) {
+      pos += header_remain_len;                   // start of keyword section
+      return read_keyword_summary(input, header_remain_len);
+
+    }).then(function(keyword_summary) {           MCommon.log(keyword_summary);
+      pos += keyword_summary.len;                 // start of key index in keyword section
+      return _slice(pos, keyword_summary.key_index_comp_len)
+                .exec(read_keyword_index, keyword_summary);
+
+    }).spread(function (keyword_summary, keyword_index) {
+      pos += keyword_summary.key_index_comp_len;  // start of keyword block in keyword section
+      slicedKeyBlock = _slice(pos, keyword_summary.key_blocks_len);
+
+      /*
+      // Now it's fast enough to look up word without key table, which scans keyword from the specified key blocks in an effcient way.
+      // No need to scan the whole key table in ahead.
+      willScanKeyTable(slicedKeyBlock, keyword_summary.num_entries, keyword_index, 00);
+      // */
+      
+      pos += keyword_summary.key_blocks_len;    // start of record section
+      
+      KEY_INDEX = keyword_index;
+      
+    }).then(function () { 
+      return _slice(pos, 32)
+                .exec(read_record_summary, pos);
+      
+    }).spread(function (record_summary) {       MCommon.log(record_summary);
+      pos += record_summary.len;                // start of record blocks in record section
+      return _slice(pos, record_summary.index_len)
+                .exec(read_record_block, record_summary);
+
+    }).spread(function() {                      MCommon.log('-- parse done --', file.name);
+      // resolve and return lookup() function according to file extension (mdx/mdd)
+      LOOKUP[ext].description = attrs.Description;
+      return resolve(LOOKUP[ext]);
+    });
+  };
+  
+  // -------------------------
+  // END OF parse_mdict()
+  // -------------------------
+  
+  /**
+   * Load a set of files which will be parsed as MDict dictionary & resource (mdx/mdd).
+   */
+  return function load(files) {
+      var resources = [];
+      Array.prototype.forEach.call(files, function(f) {
+        var ext =  MCommon.getExtension(f.name, 'mdx');
+        
+        resources.push(resources[ext] = parse_mdict(f, ext));
+      });
+        
+      return Promise.all(resources)
+                    .then(function() { return resolve(resources); });
+    };
+  
+}());
diff --git a/en/searchurl/mdict/mdict-renderer.js b/en/searchurl/mdict/mdict-renderer.js
new file mode 100644
index 00000000..18e35ecd
--- /dev/null
+++ b/en/searchurl/mdict/mdict-renderer.js
@@ -0,0 +1,200 @@
+//define(['jquery', 'bluebird', 'speex', 'pcmdata', 'bitstring'], factory);
+//$, Promise, SpeexLib, PCMDataLib
+/**
+ * Usage:
+ *   var fileList = ...; // FileList object
+ *   var word = ...;     // word for lookup
+ *   require(['mdict-parser', 'mdict-renderer'], function(MParser, MRenderer) {
+ *      MParser(fileList).then(function(resources) {
+ *         var mdict = MRenderer(resources),
+ *             dict_desc = resources.description.mdx;
+ *         mdict.lookup(word).then(function($content) {
+ *            // use $content to display result
+ *         });
+ *      });
+ *    });
+ */
+var MRenderer = (function () {
+  var MIME = {
+    'css': 'text/css',
+    'img': 'image',
+    'jpg': 'image/jpeg',
+    'png': 'image/png',
+    'spx': 'audio/x-speex',
+    'wav': 'audio/wav',
+    'mp3': 'audio/mp3',
+    'js' : 'text/javascript'
+  };
+  
+  function getExtension(filename, defaultExt) {
+    return /(?:\.([^.]+))?$/.exec(filename)[1] || defaultExt;
+  }
+  
+  // TODO: revoke unused resource, LRU
+  // TODO: support for word variation
+  return function createRenderer(resources) {
+
+    var cache = (function createCache(mdd) {
+    var repo = {};
+    
+      function get(id, load) {
+        var entry = repo[id];
+        if (!entry) {
+          repo[id] = entry = new Promise(function(resolve) {
+          var will = mdd.then(function(lookup) {
+            console.log('lookup: ' + id);
+            return lookup(id);
+          }).then(load)
+            .then(function(url) { resolve(url); });
+          });
+        }
+        return entry;
+      }
+      
+      return {get: get};
+    })(resources['mdd']);
+    
+    function loadData(mime, data) {
+      var blob = new Blob([data], {type: mime});
+      return URL.createObjectURL(blob);      
+    }
+  
+    function loadAudio(ext, data) {
+      if (ext === 'spx') {
+        var blob = decodeSpeex(String.fromCharCode.apply(null, data));
+        return URL.createObjectURL(blob);
+      } else {  // 'spx'
+        return loadData(MIME[ext] || 'audio', data);
+      }
+    }
+    
+    // TODO: LRU cache: remove oldest one only after rendering.
+    function replaceImage(index, img) {
+      var $img = $(img);
+      var src = $img.attr('src'), m = /^file:\/\/(.*)/.exec(src);
+      if (m) { src = m[1]; }
+      cache.get(src, loadData.bind(null, MIME['img']))
+           .then(function(url) {
+              $img.attr({src: url, src_: src});
+            });
+    }
+    
+    function playAudio(e, $a) {
+      ($a || $(this)).find('audio')[0].play();
+    }
+    
+    function renderAudio() {
+      var $a = $(this);
+      if ($a.attr('href_')) {
+        playAudio($a);
+      } else {
+        var href = $a.attr('href'), res = href.substring(8);
+        var ext = getExtension(res, 'wav');
+        cache.get(res, loadAudio.bind(null, ext))
+             .then(function(url) {
+                $a.append($('<audio>').attr({src: url, src_: href})).on('click', playAudio);
+                setTimeout(playAudio.bind($a));
+             });
+      }
+      return false;
+    }
+    
+    function replaceCss(index, link) {
+      var $link = $(link);
+      var href = $link.attr('href');
+      cache.get(href, loadData.bind(null, MIME['css']))
+           .then(function(url) {
+              // $link.attr({href: url, href_: href});
+              // TODO: Limit scope of embedded styles provide by mdd file
+              // TODO: use shadow dom for Chrome
+              // TODO: use scoped style for Firefox
+              $link.replaceWith($('<style scoped>', {src_: href}).text('@import url("' + url + '")'));
+            });
+    }
+    
+    function injectJS(index, el) {
+      var $el = $(el);
+      var src = $el.attr('src');
+      cache.get(src, loadData.bind(null, MIME['js']))
+           .then(function(url) {
+              $el.remove();
+              $.ajax({url: url, dataType: 'script', cache: true});
+            });
+    }    
+    
+    function decodeSpeex(file) {
+      var ogg = new Ogg(file, {file: true});
+      ogg.demux();
+
+      var header = Speex.parseHeader(ogg.frames[0]);
+      console.log(header);
+
+      var comment = new SpeexComment(ogg.frames[1]);
+      console.log(comment.data);
+
+      var spx = new Speex({
+        quality: 8,
+        mode: header.mode,
+        rate: header.rate
+      });
+      
+      var waveData = PCMData.encode({
+          sampleRate: header.rate,
+          channelCount: header.nb_channels,
+          bytesPerSample: 2,
+          data: spx.decode(ogg.bitstream(), ogg.segments)
+        });
+
+      return new Blob([Speex.util.str2ab(waveData)], {type: "audio/wav"});
+    }
+
+    function render($content) {
+      if (resources['mdd']) {
+        $content.find('img[src]').each(replaceImage);
+        
+        $content.find('link[rel=stylesheet]').each(replaceCss);
+        
+        $content.find('script[src]').each(injectJS);
+        
+        $content.find('a[href^="sound://"]').on('click', renderAudio);
+        
+        setTimeout(function() { $('#definition *').trigger('resize'); });
+      }
+      
+// resolve entry:// link dynamically in mdict.js      
+//      // rewrite in-page link
+//      $content.find('a[href^="entry://"]').each(function() { 
+//        var $el = $(this), href = $el.attr('href');
+//        if (href.match('#')) {
+//          $el.attr('href', href.substring(8));
+//        }
+//      });
+      
+      return $content;
+    }    
+    
+    return {
+      lookup: function lookup(query) {
+        return (resources['mdx'] || resources['mdd'])
+          .then(function (lookup) {
+            return lookup(query);
+          }).then(function (definitions) {
+            console.log('lookup done!');
+            var html = definitions.reduce(function(prev, txt) { 
+              return prev + '<p></p>' + txt;
+            }, '<p>' + definitions.length + ' entry(ies) </p>');
+            return Promise.resolve(render($('<div>').html(html)));
+          });
+      },
+      
+      search: function (query) {
+        return resources['mdx'].then(function(lookup) {
+          return lookup(query);
+        });
+      },
+      
+      render: render,
+    };
+  }
+
+}());
diff --git a/en/searchurl/mdict/mdict.html b/en/searchurl/mdict/mdict.html
new file mode 100644
index 00000000..5cf88375
--- /dev/null
+++ b/en/searchurl/mdict/mdict.html
@@ -0,0 +1,206 @@
+<!DOCTYPE html>
+<html lang="en-us">
+  <head>
+    <meta charset="UTF-8">
+    <title>mdict</title>
+    <meta name="viewport" content="width=device-width, initial-scale=1">
+    <link rel="stylesheet" type="text/css">
+    <link rel="stylesheet" type="text/css" href="selectize.default.css" media="screen">
+    <style>
+      #btnLookup {
+        border: none;
+        height: 36px;
+        font-size: 12pt;
+        font-weight: bold;
+        vertical-align: top;
+        border-radius: 3px;
+      }
+
+      #btnLookup:not([disabled]) {
+        background: #1A4FDD;
+        color: white;
+      }
+
+      #word + .selectize-control {
+        display: inline-block;
+        min-width: 18em;
+      }
+    </style>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/npm/jquery@3.6.1/dist/jquery.min.js"></script>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/gh/fengdh/mdict-js/selectize.min.js"></script>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/gh/nodeca/pako/dist/pako_inflate.min.js"></script>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/npm/bluebird@3.7.2/js/browser/bluebird.min.js"></script>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/gh/torappinfo/uweb/en/searchurl/mdict/ripemd128.min.js"></script>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/gh/torappinfo/uweb/en/searchurl/mdict/mdict-common.min.js"></script>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/gh/torappinfo/uweb/en/searchurl/mdict/mdict-parser.min.js"></script>
+    <script type="text/javascript" src="https://cdn.jsdelivr.net/gh/torappinfo/uweb/en/searchurl/mdict/mdict-renderer.min.js"></script>
+  </head>
+  <body>
+        Choose a dictionary file (*.mdx + optional *.mdd): <input id="dictfile" type="file" multiple>
+        <p>
+        <input id="word" type="text" value="">
+        <input id="btnLookup" type="button" value="look up" disabled="false">
+        <div id="definition">
+        </div>
+      
+    <script>
+      //file:///...mdictstem?app=....html#word=
+      var dictinput = document.getElementById('dictfile');
+      var wordinput = document.getElementById('word');
+      var btnLookup = document.getElementById('btnLookup');
+      var definition = document.getElementById('definition');
+
+      window.onhashchange = function() {
+        let v = location.hash.substring(6);
+        if(v.length>0){
+          wordinput.value = v;
+          //wordinput.dispatchEvent(new Event('change', {'bubbles': true}));
+          btnLookup.click();
+        }
+      };
+
+      {
+        let v = decodeURIComponent(location.hash.substring(6));
+        wordinput.value = v;
+      }
+
+      $(wordinput).selectize({maxItems: 1});
+      function accept(e) {
+        var fileList = $(e.target).prop('files');
+        $(btnLookup).attr('disabled', true);
+        if (fileList.length > 0) {
+          $(btnLookup).addClass('stripes');
+          $(wordinput).on('keyup', function(e) { e.which === 13
+                                               && $(btnLookup).click(); });
+          MParser(fileList).then(function(resources) {
+            var mdict = MRenderer(resources);
+            function doSearch(phrase, offset) {
+              console.log(phrase + '');
+              mdict.lookup(phrase, offset).then(function($content) {
+                $(definition).empty().append($content.contents());
+                console.log('--');
+              });
+            }
+
+            $(btnLookup)
+              .attr('disabled', false)
+              .off('.#mdict')
+              .on('click.#mdict', function() {
+                doSearch($(wordinput).val());
+              }).click();
+
+            $(wordinput)[0].selectize.destroy();
+
+            $(wordinput).selectize({
+              plugins: ['restore_on_backspace'],
+              maxItems: 1,
+              maxOptions: 1 << 20,
+              valueField: 'value',
+              labelField: 'word',
+              searchField: 'word',
+              delimiter: '~~',
+              loadThrottle: 10,
+              create: function(v, callback) {
+                return callback({word: v, value: v});
+              },
+              createOnBlur: true,
+              closeAfterSelect: true,
+              allowEmptyOption: true,
+              score: function(search) {
+                var score =
+                    this.getScoreFunction(search);
+                return function(item) {
+                  return 1;
+                };
+              },
+              load: function(query, callback) {
+                var self = this;
+                if (!query.length) {
+                  this.clearOptions();
+                  this.refreshOptions();
+                  return;
+                };
+
+                mdict.search({phrase: query, max: 5000}).then(function(list) {
+                  //                  console.log(list.join(', '));
+                  // TODO: filter candidate keyword starting with "_"
+                  list = list.map(function(v) {
+                    return {word: v, value: v.offset};
+                  });
+                  self.clearOptions();
+                  callback(list);
+                });
+              },
+              onChange: function(value) {
+                var item = this.options[value];
+                if (item) {
+                  var value = item.word;
+                  doSearch(value, value.offset);
+                  $(wordinput).val(value);
+                } else {
+                  $(definition).empty();
+                }
+              },
+            });
+          }).catch(err => alert(err)) ;
+        } else {
+          $(btnLookup).attr('disabled', false);
+        }
+
+        // jump to word with link started with "entry://"
+        // TODO: have to ignore in-page jump
+        $(definition).on('click', 'a', function(e) {
+          var href = $(this).attr('href');
+          if (href && href.substring(0, 8) === 'entry://') {
+            var word = href.substring(8);
+            // TODO: remove '#' to get jump target
+            if (word.charAt(0) !== '#') {
+              word = word.replace(/(^[/\\])|([/]$)/, '');
+
+              $(wordinput).val(word);
+              $(btnLookup).click();
+            } else {
+              var currentUrl = location.href;
+              location.href = word;                       //Go to the target element.
+              history.replaceState(null,null,currentUrl);
+            }
+            return false;
+          }
+        });
+      }
+      $(dictinput).on('change', accept);
+ 
+      {
+        let path = location.pathname;
+        let iSlash = path.lastIndexOf('/');
+        let filenames=path.substring(iSlash+1);
+        {
+          let iDot = filenames.indexOf('.');
+          if(iDot>0)
+            document.getElementsByTagName("link")[0].href =
+              filenames.substring(0,iDot)+"css";
+        }
+        window.onload = function(){
+        if(filenames.length>0){
+          let clickurl = "i:5fdictinput.click():"+filenames;
+          location.href=clickurl;
+       
+          var callbackTimer = setInterval(function() {
+            let files = dictinput.files;
+            if(files.length>0){
+              if(!btnLookup.disabled){
+                clearInterval(callbackTimer);
+                return;
+              }
+            }else
+              location.href=clickurl;
+            dictinput.dispatchEvent(new Event('change', {'bubbles': true}));
+          }, 100);
+        
+        }
+        };
+      }
+      
+    </script>
+  </body>
+</html>
diff --git a/en/searchurl/mdict/ripemd128.js b/en/searchurl/mdict/ripemd128.js
new file mode 100644
index 00000000..f961ec0d
--- /dev/null
+++ b/en/searchurl/mdict/ripemd128.js
@@ -0,0 +1,135 @@
+/*
+ * A pure JavaScript implementation of RIPEMD128 using Uint8Array as input/output.
+ * By Feng Dihai <fengdh@gmail.com>, 2015/07/09
+ *
+ * Based on coiscir/jsdigest (https://github.com/coiscir/jsdigest/blob/master/src/hash/ripemd128.js)
+ * 
+ * ripemd128.js is free software released under terms of the MIT License.
+ * You can get a copy on http://opensource.org/licenses/MIT.
+ * 
+ *
+ * RIPEMD-128 (c) 1996 Hans Dobbertin, Antoon Bosselaers, and Bart Preneel
+ */
+var ripemd128;
+(function() {
+  // implementation
+  
+  // convert array of number to Uint32Array
+  function asUint32Array(arr) {
+    return new Uint32Array(arr);
+  }
+  
+  // concat 2 typed array
+  function concat(a, b) {
+    var c = new a.constructor(a.length + b.length);
+    c.set(a);
+    c.set(b, a.length);
+    return c;
+  }
+  
+  // swap high and low bits of a 32-bit int.
+  function rotl( x, n ) {
+    return ( x >>> ( 32 - n ) ) | ( x << n );
+  }
+  
+  var DIGEST = 128,
+      BLOCK = 64,
+      S = [
+        [ 11, 14, 15, 12,  5,  8,  7,  9, 11, 13, 14, 15,  6,  7,  9,  8 ], // round 1
+        [  7,  6,  8, 13, 11,  9,  7, 15,  7, 12, 15,  9, 11,  7, 13, 12 ], // round 2
+        [ 11, 13,  6,  7, 14,  9, 13, 15, 14,  8, 13,  6,  5, 12,  7,  5 ], // round 3
+        [ 11, 12, 14, 15, 14, 15,  9,  8,  9, 14,  5,  6,  8,  6,  5, 12 ], // round 4
+        [  8,  9,  9, 11, 13, 15, 15,  5,  7,  7,  8, 11, 14, 14, 12,  6 ], // parallel round 1
+        [  9, 13, 15,  7, 12,  8,  9, 11,  7,  7, 12,  7,  6, 15, 13, 11 ], // parallel round 2
+        [  9,  7, 15, 11,  8,  6,  6, 14, 12, 13,  5, 14, 13, 13,  7,  5 ], // parallel round 3
+        [ 15,  5,  8, 11, 14, 14,  6, 14,  6,  9, 12,  9, 12,  5, 15,  8 ]  // parallel round 4
+      ].map(asUint32Array),
+      X = [
+        [  0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15 ], // round 1
+        [  7,  4, 13,  1, 10,  6, 15,  3, 12,  0,  9,  5,  2, 14, 11,  8 ], // round 2
+        [  3, 10, 14,  4,  9, 15,  8,  1,  2,  7,  0,  6, 13, 11,  5, 12 ], // round 3
+        [  1,  9, 11, 10,  0,  8, 12,  4, 13,  3,  7, 15, 14,  5,  6,  2 ], // round 4
+        [  5, 14,  7,  0,  9,  2, 11,  4, 13,  6, 15,  8,  1, 10,  3, 12 ], // parallel round 1
+        [  6, 11,  3,  7,  0, 13,  5, 10, 14, 15,  8, 12,  4,  9,  1,  2 ], // parallel round 2
+        [ 15,  5,  1,  3,  7, 14,  6,  9, 11,  8, 12,  2, 10,  0,  4, 13 ], // parallel round 3
+        [  8,  6,  4,  1,  3, 11, 15,  0,  5, 12,  2, 13,  9,  7, 10, 14 ]  // parallel round 4
+      ].map(asUint32Array),
+      K = asUint32Array([
+        0x00000000, // FF
+        0x5a827999, // GG
+        0x6ed9eba1, // HH
+        0x8f1bbcdc, // II
+        0x50a28be6, // III
+        0x5c4dd124, // HHH
+        0x6d703ef3, // GGG
+        0x00000000  // FFF
+      ]),
+      F = [
+        function ( x, y, z ) {
+          return ( x ^ y ^ z );
+        },
+        function ( x, y, z ) {
+          return ( x & y ) | ( ( ~x ) & z );
+        },
+        function ( x, y, z ) {
+          return ( x | ( ~y ) ) ^ z;
+        },
+        function ( x, y, z ) {
+          return ( x & z ) | ( y & ( ~z ) );
+        }
+      ];
+
+  ripemd128 = function( data ) {
+    var aa, bb, cc, dd, aaa, bbb, ccc, ddd, i, l, r, rr, t, tmp, x,
+        hash = new Uint32Array([ 0x67452301, 0xefcdab89, 0x98badcfe, 0x10325476 ]),
+        bytes = data.length;
+    
+    var padding = new Uint8Array(( ( bytes % 64 ) < 56 ? 56 : 120 ) - ( bytes % 64 ));
+    padding[0] = [0x80];
+    
+    data = new Uint32Array(concat(data, padding).buffer);
+    
+    // ending with check bits (= little endian 64-bit int, 8 * data.length)
+    bytes = bytes << 3;
+    x = concat( data, [bytes, bytes >> 31 >> 1]);
+    // update hash
+    for ( i = 0, t = 0, l = x.length; i < l; i += 16, t = 0 ) {
+      aa = aaa = hash[0];
+      bb = bbb = hash[1];
+      cc = ccc = hash[2];
+      dd = ddd = hash[3];
+      
+      for ( ; t < 64; ++t ) {
+        r   = ~~( t / 16 );
+        aa  = rotl(aa + F[r]( bb, cc, dd ) + x[ i + X[r][ t % 16 ] ] + K[r], S[r][ t % 16 ]);
+        
+        tmp = dd;
+        dd  = cc;
+        cc  = bb;
+        bb  = aa;
+        aa  = tmp;
+      }
+      
+      for ( ; t < 128; ++t ) {
+        r   = ~~( t / 16 );
+        rr  = ~~( ( 63 - ( t % 64 ) ) / 16 );
+        aaa = rotl(aaa + F[rr]( bbb, ccc, ddd ) + x[ i + X[r][ t % 16 ] ] + K[r], S[r][ t % 16 ]);
+        
+        tmp = ddd;
+        ddd = ccc;
+        ccc = bbb;
+        bbb = aaa;
+        aaa = tmp;
+      }
+      
+      ddd     = hash[1] + cc + ddd;
+      hash[1] = hash[2] + dd + aaa;
+      hash[2] = hash[3] + aa + bbb;
+      hash[3] = hash[0] + bb + ccc;
+      hash[0] = ddd;
+    }
+    
+    return new Uint8Array( hash.buffer );
+  }
+  
+}());
diff --git a/en/searchurl/mdict/selectize.default.css b/en/searchurl/mdict/selectize.default.css
new file mode 100644
index 00000000..ffd219c4
--- /dev/null
+++ b/en/searchurl/mdict/selectize.default.css
@@ -0,0 +1,387 @@
+/**
+ * selectize.default.css (v0.12.1) - Default Theme
+ * Copyright (c) 2013–2015 Brian Reavis & contributors
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License"); you may not use this
+ * file except in compliance with the License. You may obtain a copy of the License at:
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software distributed under
+ * the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF
+ * ANY KIND, either express or implied. See the License for the specific language
+ * governing permissions and limitations under the License.
+ *
+ * @author Brian Reavis <brian@thirdroute.com>
+ */
+.selectize-control.plugin-drag_drop.multi > .selectize-input > div.ui-sortable-placeholder {
+  visibility: visible !important;
+  background: #f2f2f2 !important;
+  background: rgba(0, 0, 0, 0.06) !important;
+  border: 0 none !important;
+  -webkit-box-shadow: inset 0 0 12px 4px #ffffff;
+  box-shadow: inset 0 0 12px 4px #ffffff;
+}
+.selectize-control.plugin-drag_drop .ui-sortable-placeholder::after {
+  content: '!';
+  visibility: hidden;
+}
+.selectize-control.plugin-drag_drop .ui-sortable-helper {
+  -webkit-box-shadow: 0 2px 5px rgba(0, 0, 0, 0.2);
+  box-shadow: 0 2px 5px rgba(0, 0, 0, 0.2);
+}
+.selectize-dropdown-header {
+  position: relative;
+  padding: 5px 8px;
+  border-bottom: 1px solid #d0d0d0;
+  background: #f8f8f8;
+  -webkit-border-radius: 3px 3px 0 0;
+  -moz-border-radius: 3px 3px 0 0;
+  border-radius: 3px 3px 0 0;
+}
+.selectize-dropdown-header-close {
+  position: absolute;
+  right: 8px;
+  top: 50%;
+  color: #303030;
+  opacity: 0.4;
+  margin-top: -12px;
+  line-height: 20px;
+  font-size: 20px !important;
+}
+.selectize-dropdown-header-close:hover {
+  color: #000000;
+}
+.selectize-dropdown.plugin-optgroup_columns .optgroup {
+  border-right: 1px solid #f2f2f2;
+  border-top: 0 none;
+  float: left;
+  -webkit-box-sizing: border-box;
+  -moz-box-sizing: border-box;
+  box-sizing: border-box;
+}
+.selectize-dropdown.plugin-optgroup_columns .optgroup:last-child {
+  border-right: 0 none;
+}
+.selectize-dropdown.plugin-optgroup_columns .optgroup:before {
+  display: none;
+}
+.selectize-dropdown.plugin-optgroup_columns .optgroup-header {
+  border-top: 0 none;
+}
+.selectize-control.plugin-remove_button [data-value] {
+  position: relative;
+  padding-right: 24px !important;
+}
+.selectize-control.plugin-remove_button [data-value] .remove {
+  z-index: 1;
+  /* fixes ie bug (see #392) */
+  position: absolute;
+  top: 0;
+  right: 0;
+  bottom: 0;
+  width: 17px;
+  text-align: center;
+  font-weight: bold;
+  font-size: 12px;
+  color: inherit;
+  text-decoration: none;
+  vertical-align: middle;
+  display: inline-block;
+  padding: 2px 0 0 0;
+  border-left: 1px solid #0073bb;
+  -webkit-border-radius: 0 2px 2px 0;
+  -moz-border-radius: 0 2px 2px 0;
+  border-radius: 0 2px 2px 0;
+  -webkit-box-sizing: border-box;
+  -moz-box-sizing: border-box;
+  box-sizing: border-box;
+}
+.selectize-control.plugin-remove_button [data-value] .remove:hover {
+  background: rgba(0, 0, 0, 0.05);
+}
+.selectize-control.plugin-remove_button [data-value].active .remove {
+  border-left-color: #00578d;
+}
+.selectize-control.plugin-remove_button .disabled [data-value] .remove:hover {
+  background: none;
+}
+.selectize-control.plugin-remove_button .disabled [data-value] .remove {
+  border-left-color: #aaaaaa;
+}
+.selectize-control {
+  position: relative;
+}
+.selectize-dropdown,
+.selectize-input,
+.selectize-input input {
+  color: #303030;
+  font-family: inherit;
+  font-size: 13px;
+  line-height: 18px;
+  -webkit-font-smoothing: inherit;
+}
+.selectize-input,
+.selectize-control.single .selectize-input.input-active {
+  background: #ffffff;
+  cursor: text;
+  display: inline-block;
+}
+.selectize-input {
+  border: 1px solid #d0d0d0;
+  padding: 8px 8px;
+  display: inline-block;
+  width: 100%;
+  overflow: hidden;
+  position: relative;
+  z-index: 1;
+  -webkit-box-sizing: border-box;
+  -moz-box-sizing: border-box;
+  box-sizing: border-box;
+  -webkit-box-shadow: inset 0 1px 1px rgba(0, 0, 0, 0.1);
+  box-shadow: inset 0 1px 1px rgba(0, 0, 0, 0.1);
+  -webkit-border-radius: 3px;
+  -moz-border-radius: 3px;
+  border-radius: 3px;
+}
+.selectize-control.multi .selectize-input.has-items {
+  padding: 5px 8px 2px;
+}
+.selectize-input.full {
+  background-color: #ffffff;
+}
+.selectize-input.disabled,
+.selectize-input.disabled * {
+  cursor: default !important;
+}
+.selectize-input.focus {
+  -webkit-box-shadow: inset 0 1px 2px rgba(0, 0, 0, 0.15);
+  box-shadow: inset 0 1px 2px rgba(0, 0, 0, 0.15);
+}
+.selectize-input.dropdown-active {
+  -webkit-border-radius: 3px 3px 0 0;
+  -moz-border-radius: 3px 3px 0 0;
+  border-radius: 3px 3px 0 0;
+}
+.selectize-input > * {
+  vertical-align: baseline;
+  display: -moz-inline-stack;
+  display: inline-block;
+  zoom: 1;
+  *display: inline;
+}
+.selectize-control.multi .selectize-input > div {
+  cursor: pointer;
+  margin: 0 3px 3px 0;
+  padding: 2px 6px;
+  background: #1da7ee;
+  color: #ffffff;
+  border: 1px solid #0073bb;
+}
+.selectize-control.multi .selectize-input > div.active {
+  background: #92c836;
+  color: #ffffff;
+  border: 1px solid #00578d;
+}
+.selectize-control.multi .selectize-input.disabled > div,
+.selectize-control.multi .selectize-input.disabled > div.active {
+  color: #ffffff;
+  background: #d2d2d2;
+  border: 1px solid #aaaaaa;
+}
+.selectize-input > input {
+  display: inline-block !important;
+  padding: 0 !important;
+  min-height: 0 !important;
+  max-height: none !important;
+  max-width: 100% !important;
+  margin: 0 1px !important;
+  text-indent: 0 !important;
+  border: 0 none !important;
+  background: none !important;
+  line-height: inherit !important;
+  -webkit-user-select: auto !important;
+  -webkit-box-shadow: none !important;
+  box-shadow: none !important;
+}
+.selectize-input > input::-ms-clear {
+  display: none;
+}
+.selectize-input > input:focus {
+  outline: none !important;
+}
+.selectize-input::after {
+  content: ' ';
+  display: block;
+  clear: left;
+}
+.selectize-input.dropdown-active::before {
+  content: ' ';
+  display: block;
+  position: absolute;
+  background: #f0f0f0;
+  height: 1px;
+  bottom: 0;
+  left: 0;
+  right: 0;
+}
+.selectize-dropdown {
+  position: absolute;
+  z-index: 10;
+  border: 1px solid #d0d0d0;
+  background: #ffffff;
+  margin: -1px 0 0 0;
+  border-top: 0 none;
+  -webkit-box-sizing: border-box;
+  -moz-box-sizing: border-box;
+  box-sizing: border-box;
+  -webkit-box-shadow: 0 1px 3px rgba(0, 0, 0, 0.1);
+  box-shadow: 0 1px 3px rgba(0, 0, 0, 0.1);
+  -webkit-border-radius: 0 0 3px 3px;
+  -moz-border-radius: 0 0 3px 3px;
+  border-radius: 0 0 3px 3px;
+}
+.selectize-dropdown [data-selectable] {
+  cursor: pointer;
+  overflow: hidden;
+}
+.selectize-dropdown [data-selectable] .highlight {
+  background: rgba(125, 168, 208, 0.2);
+  -webkit-border-radius: 1px;
+  -moz-border-radius: 1px;
+  border-radius: 1px;
+}
+.selectize-dropdown [data-selectable],
+.selectize-dropdown .optgroup-header {
+  padding: 5px 8px;
+}
+.selectize-dropdown .optgroup:first-child .optgroup-header {
+  border-top: 0 none;
+}
+.selectize-dropdown .optgroup-header {
+  color: #303030;
+  background: #ffffff;
+  cursor: default;
+}
+.selectize-dropdown .active {
+  background-color: #f5fafd;
+  color: #495c68;
+}
+.selectize-dropdown .active.create {
+  color: #495c68;
+}
+.selectize-dropdown .create {
+  color: rgba(48, 48, 48, 0.5);
+}
+.selectize-dropdown-content {
+  overflow-y: auto;
+  overflow-x: hidden;
+  max-height: 200px;
+}
+.selectize-control.single .selectize-input,
+.selectize-control.single .selectize-input input {
+  cursor: pointer;
+}
+.selectize-control.single .selectize-input.input-active,
+.selectize-control.single .selectize-input.input-active input {
+  cursor: text;
+}
+.selectize-control.single .selectize-input:after {
+  content: ' ';
+  display: block;
+  position: absolute;
+  top: 50%;
+  right: 15px;
+  margin-top: -3px;
+  width: 0;
+  height: 0;
+  border-style: solid;
+  border-width: 5px 5px 0 5px;
+  border-color: #808080 transparent transparent transparent;
+}
+.selectize-control.single .selectize-input.dropdown-active:after {
+  margin-top: -4px;
+  border-width: 0 5px 5px 5px;
+  border-color: transparent transparent #808080 transparent;
+}
+.selectize-control.rtl.single .selectize-input:after {
+  left: 15px;
+  right: auto;
+}
+.selectize-control.rtl .selectize-input > input {
+  margin: 0 4px 0 -2px !important;
+}
+.selectize-control .selectize-input.disabled {
+  opacity: 0.5;
+  background-color: #fafafa;
+}
+.selectize-control.multi .selectize-input.has-items {
+  padding-left: 5px;
+  padding-right: 5px;
+}
+.selectize-control.multi .selectize-input.disabled [data-value] {
+  color: #999;
+  text-shadow: none;
+  background: none;
+  -webkit-box-shadow: none;
+  box-shadow: none;
+}
+.selectize-control.multi .selectize-input.disabled [data-value],
+.selectize-control.multi .selectize-input.disabled [data-value] .remove {
+  border-color: #e6e6e6;
+}
+.selectize-control.multi .selectize-input.disabled [data-value] .remove {
+  background: none;
+}
+.selectize-control.multi .selectize-input [data-value] {
+  text-shadow: 0 1px 0 rgba(0, 51, 83, 0.3);
+  -webkit-border-radius: 3px;
+  -moz-border-radius: 3px;
+  border-radius: 3px;
+  background-color: #1b9dec;
+  background-image: -moz-linear-gradient(top, #1da7ee, #178ee9);
+  background-image: -webkit-gradient(linear, 0 0, 0 100%, from(#1da7ee), to(#178ee9));
+  background-image: -webkit-linear-gradient(top, #1da7ee, #178ee9);
+  background-image: -o-linear-gradient(top, #1da7ee, #178ee9);
+  background-image: linear-gradient(to bottom, #1da7ee, #178ee9);
+  background-repeat: repeat-x;
+  filter: progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff1da7ee', endColorstr='#ff178ee9', GradientType=0);
+  -webkit-box-shadow: 0 1px 0 rgba(0,0,0,0.2),inset 0 1px rgba(255,255,255,0.03);
+  box-shadow: 0 1px 0 rgba(0,0,0,0.2),inset 0 1px rgba(255,255,255,0.03);
+}
+.selectize-control.multi .selectize-input [data-value].active {
+  background-color: #0085d4;
+  background-image: -moz-linear-gradient(top, #008fd8, #0075cf);
+  background-image: -webkit-gradient(linear, 0 0, 0 100%, from(#008fd8), to(#0075cf));
+  background-image: -webkit-linear-gradient(top, #008fd8, #0075cf);
+  background-image: -o-linear-gradient(top, #008fd8, #0075cf);
+  background-image: linear-gradient(to bottom, #008fd8, #0075cf);
+  background-repeat: repeat-x;
+  filter: progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff008fd8', endColorstr='#ff0075cf', GradientType=0);
+}
+.selectize-control.single .selectize-input {
+  -webkit-box-shadow: 0 1px 0 rgba(0,0,0,0.05), inset 0 1px 0 rgba(255,255,255,0.8);
+  box-shadow: 0 1px 0 rgba(0,0,0,0.05), inset 0 1px 0 rgba(255,255,255,0.8);
+  background-color: #f9f9f9;
+  background-image: -moz-linear-gradient(top, #fefefe, #f2f2f2);
+  background-image: -webkit-gradient(linear, 0 0, 0 100%, from(#fefefe), to(#f2f2f2));
+  background-image: -webkit-linear-gradient(top, #fefefe, #f2f2f2);
+  background-image: -o-linear-gradient(top, #fefefe, #f2f2f2);
+  background-image: linear-gradient(to bottom, #fefefe, #f2f2f2);
+  background-repeat: repeat-x;
+  filter: progid:DXImageTransform.Microsoft.gradient(startColorstr='#fffefefe', endColorstr='#fff2f2f2', GradientType=0);
+}
+.selectize-control.single .selectize-input,
+.selectize-dropdown.single {
+  border-color: #b8b8b8;
+}
+.selectize-dropdown .optgroup-header {
+  padding-top: 7px;
+  font-weight: bold;
+  font-size: 0.85em;
+}
+.selectize-dropdown .optgroup {
+  border-top: 1px solid #f0f0f0;
+}
+.selectize-dropdown .optgroup:first-child {
+  border-top: 0 none;
+}
diff --git a/index.html b/index.html
index db7d507e..33185bf8 100644
--- a/index.html
+++ b/index.html
@@ -1 +1,10 @@
-<meta http-equiv="refresh" content="0; url=en/index.html">
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <title>en/</title>
+    <link rel="canonical" href="en/">
+    <meta name="robots" content="noindex">
+    <meta charset="utf-8">
+    <meta http-equiv="refresh" content="0; url=en/">
+  </head>
+</html>
diff --git a/zh/index.html b/zh/index.html
index 30430013..3d86020f 100644
--- a/zh/index.html
+++ b/zh/index.html
@@ -1,7 +1,7 @@
 <!doctype html>
 <html>
   <head>
-	<meta name="generator" content="Hugo 0.104.2" />
+	<meta name="generator" content="Hugo 0.104.3" />
     <meta charset="utf-8">
     <meta name="viewport" content="width=device-width, initial-scale=1" />
     
diff --git a/zh/searchurl/txt/mdict.html b/zh/searchurl/txt/mdict.html
index b29150b9..94399516 100644
--- a/zh/searchurl/txt/mdict.html
+++ b/zh/searchurl/txt/mdict.html
@@ -22,31 +22,7 @@
       }
 
       #dict-title {
-        position: fixed;
-        bottom: 0;
-        right: 0;
-        max-width: 300px;
-        font-size: 10px;
-        opacity: 0.9;
-        background: #DDD;
-        box-shadow: -2px -2px 4px 4px rgba(0, 0, 0, 0.3);
-      }
-
-      #dict-title * {
-        font-size: 10px!important;
-      }
-      
-      #mdict-online-viewer {
-        font-size: 14px;
-        font-family: "Georgia", "Times New Roman";
-        height: 100%;
-        width: 100%;
-      }
-      
-      #mdict-online-viewer #definition {
-        font-size: 14px;
-        height: 100%;
-        overflow: auto;
+        display:none;
       }
       
       #word + .selectize-control {
@@ -59,9 +35,6 @@
     <script src="require.js" data-main="mdict"></script>
   </head>
   <body>
-    <section class="main-content">
-
-      <div id="mdict-online-viewer">
         Choose a dictionary file (*.mdx + optional *.mdd): <input id="dictfile" type="file" multiple>
         <p>
         <input id="word" type="text" value="">
@@ -70,9 +43,6 @@
         <div id="dict-title"></div>
         <div id="definition">
         </div>
-      </div>
-      
-    </section>
     <script>
       //?file=.../xx.xx,.../xx.xxx#word=
       var dictinput = document.getElementById('dictfile');
@@ -97,7 +67,6 @@
         let filenames = location.search.substring(6);
         window.onload = function(){
         if(filenames.length>0){
-          filenames = decodeURIComponent(filenames);
           let clickurl = "i:5fdictinput.click():"+filenames;
           location.href=clickurl;
           //setTimeout(()=>{dictinput.dispatchEvent(new Event('change', {'bubbles': true}));},100);
diff --git a/zh/searchurl/txt/txtview.html b/zh/searchurl/txt/txtview.html
index 51efb255..41baa37b 100644
--- a/zh/searchurl/txt/txtview.html
+++ b/zh/searchurl/txt/txtview.html
@@ -21,7 +21,7 @@
     }
     
     var file;
-    var blocksize = 1024;
+    var blocksize = 20480;
     function loadSlice(){
       function render(u8array){
         document.body.innerHTML = marked(new TextDecoder().decode(u8array));
