diff --git a/README.md b/README.md
index 375a27e..15651bd 100644
--- a/README.md
+++ b/README.md
@@ -35,12 +35,14 @@ The following examples where given after the system was collecting for about one
 
 * install npm and node if you have not already (http://howtonode.org/introduction-to-npm or http://nodejs.org/)
 * install / start your redis server (http://redis.io/topics/quickstart) on a disk with several free GB
+
 * clone this repo "git clone https://github.com/monbro/opensemanticapi.git"
 * change config if needed in "/config.js"
+* NOTE: depending on what you want (scraper / cronjob to run or http api server, set 'http_server' to true or false)
 * open the repository folder in your console
 * enter "npm install", it will install all dependencies automatically
 * start the node server with "node app.js"
-* now it should print what it is collecting
+* now it should print what it is collecting or what http route is requested
 * the longer it collects data the better the results should be
 * now you can access the relations through your browser like http://localhost:8080/relations/database or direct by accessing your redis server
 
diff --git a/app.js b/app.js
index 2aeb5e1..3f23af1 100644
--- a/app.js
+++ b/app.js
@@ -9,73 +9,44 @@
  * @version 0.1
  */
 
-/* Used variables names
- * ____sites____ = wikipedia page titles which were scraped already
- * ____sites2do____ = wikipedia page titles which are queued to scrape
- * ____all____ = a collection of all ever seen words with a increment number
- */
-
 /** 
- * thrid part modules
+ * laod config
  */
-var restify = require('restify');
-var express = require('express');
 var config = require('./config');
-var tools = require('./lib/tools');
-var redis = require("redis");
-// var mongoose = require('mongoose');
-var _ = require("underscore");
-var $ = require("jquery");
 
 /** 
  * Basic Objects
  */
 
-// create redis client
-var client = redis.createClient();
-
-/** 
- * our classes
- */
-var text = require("./classes/textanalyze")(client);
-var wiki = require("./classes/wiki")(client);
-var apiprovider = require("./classes/apiprovider")(client);
-
-/** 
- * Objects
- */
-
+// our basic app object
+var App = function() {};
 
+// our basic app
+var app = new App();
 
-// create restify server to server http api
-var server = restify.createServer();
-server.use(restify.bodyParser());
+var Model = require('./app/model');
 
-// create restify json client for api requests
-var wikipedia = restify.createJsonClient({
-  url: 'http://'+config.creds.lang+'.wikipedia.org',
-  version: '*'
-});
+App.prototype.getModel = function(s) {
+  return new Model(s); 
+};
 
-/** 
- * Run
- */
+// var test = app.getModel('test');
+// console.log(test.getValue());
 
-// start api requests with given keyword
-// wikiSearch('database'); // database can be replaced with a random name to start with
-
-/** 
- * Routes
- */
+if(!config.creds.http_server) {
+    var Scraper = require("./app/scraping");
+    var scraper = new Scraper();
 
-// Set up our routes
-server.get('/relations/:name', apiprovider.getRelations);
+    // Start Cronjob
+    scraper.wikiSearch('database');
+}
 
 /** 
  * Server
  */
 
-// start the server
-server.listen(config.creds.server_port, function() {
-  console.log('%s listening at %s', server.name, server.url);
-});
\ No newline at end of file
+// Start HTTP API RESTFUL Server
+if(config.creds.http_server) {
+    var Http = require("./app/http");
+    var http = new Http(); 
+}
diff --git a/app/analyse.js b/app/analyse.js
new file mode 100644
index 0000000..e058a52
--- /dev/null
+++ b/app/analyse.js
@@ -0,0 +1,84 @@
+var Tools = require('../lib/tools');
+var Db = require('./db');
+var config = require('../config.js');
+var $ = require("jquery");
+
+var Analyse = function() {
+    tools = Tools;
+    // tools = new Tools();
+    db = new Db();
+};
+
+Analyse.prototype.scanTextBlock = function(snippet,counter) {
+    if(config.creds.debug) {
+        console.log('scanTextBlock');
+    }
+    // split the text block to words
+    var words = tools.tokenize(snippet);
+
+    if(config.creds.debug) {
+        console.log('Count of words in snippet ('+counter+'): '+words.length);
+    }
+
+      // create empty object
+      var obj = {};
+      // var multi = client.multi();
+      db.enableMulti();
+
+      // loop all words
+      for (var i = words.length - 1; i >= 0; i--) {
+
+        // count all seen words
+        if(typeof obj[words[i].toLowerCase()] == 'undefined')
+          obj[words[i].toLowerCase()] = 1;
+        else
+          obj[words[i].toLowerCase()]++;
+
+        // add every word to the queue to spread the scrape
+        // multi.sadd('____sites2do____',words[i].toLowerCase());
+        db.addPageToQueue(words[i].toLowerCase());
+
+        // if(config.creds.debug)
+        //   console.log(words[i].toLowerCase()+'¥ - '+words[j].toLowerCase()+' - '+similar_text(words[i].toLowerCase(),words[j].toLowerCase(),1));
+      }
+
+      // var base;
+
+      $.each(obj, function(index, value) {
+
+        // skip if not valid
+        if(typeof index == 'undefined' || typeof index.toLowerCase == 'undefined')
+          return;
+
+        // create new obj from class Base, make sure to work with lowercase only
+        // base = new Base(index.toLowerCase());
+
+        // loop all words
+        $.each(obj, function(index2, value2) {
+          if(index != index2) {
+            // add relation, value2 is the counter of how often the word was seen in the recent textblock
+            // base.pushRelation(index2.toLowerCase(),value2);
+            db.addRelation(index.toLowerCase(),index2.toLowerCase(),value2);
+          }
+        });
+
+        // base.save();
+
+        // add to our general 'ALL' collection, to identify the most used words of all
+        // multi.sadd('____all____', index.toLowerCase()); // add keyword
+        // multi.incrby('____all____'+':'+index.toLowerCase(), value); // track its density
+        db.addToGlobalCounter(value,index.toLowerCase());
+
+      });
+
+    // flush changes to database
+    db.flush(function(err, replies) {
+        return true;
+    });
+
+      // multi.exec(function(err, replies) {
+      //     return true;
+      // });
+};
+
+module.exports = Analyse;
\ No newline at end of file
diff --git a/app/db.js b/app/db.js
new file mode 100644
index 0000000..32aa4f3
--- /dev/null
+++ b/app/db.js
@@ -0,0 +1,110 @@
+var Tools = require('../lib/tools');
+var config = require('../config.js');
+var redis = require("redis");
+
+/* Used variables names
+ * ____sites____ = wikipedia page titles which were scraped already
+ * ____sites2do____ = wikipedia page titles which are queued to scrape
+ * ____all____ = a collection of all ever seen words with a increment number
+ */
+
+var Db = function() {
+    // create redis client
+    tools = Tools;
+    that = this;
+    client = redis.createClient();
+    isMulti = false;
+};
+
+Db.prototype.enableMulti = function() {
+    multi = client.multi();
+    isMulti = true;
+};
+
+Db.prototype.disableMulti = function() {
+    multi = null;
+    isMulti = false;
+};
+
+Db.prototype.addPageToQueue = function(s) {
+    if(config.creds.debug) {
+        console.log('addPageToQueue "'+s+'"');
+    }
+    if(isMulti) {
+        multi.sadd('____sites2do____',s);
+    }
+    else {
+        client.sadd('<',s);
+    }
+};
+
+Db.prototype.removePageFromQueue = function(s) {
+    client.srem('____sites2do____',s);
+};
+
+Db.prototype.getRandomItemFromQueue = function(callback) {
+    client.srandmember('____sites2do____', function (err, result) {
+        return callback(result);
+    });
+};
+
+Db.prototype.addPageAsDone = function(s, callback) {
+    client.sadd('____sites____', s, callback);
+};
+
+Db.prototype.removePageAsDone = function(s, callback) {
+    client.srem('____sites____', s, callback);
+};
+
+Db.prototype.addRelation = function(owner,relation,i) {
+    if(isMulti) {
+        multi.sadd(owner, relation);
+        if(typeof i == 'undefined') {
+          i = 1;
+        }
+        multi.incrby(owner+':'+relation, i);
+    }
+    else {
+        client.sadd(owner, relation);
+        if(typeof i == 'undefined') {
+          i = 1;
+        }
+        client.incrby(owner+':'+relation, i);
+    }
+};
+
+Db.prototype.addToGlobalCounter = function(owner,s) {
+    if(isMulti) {
+        // add to our general 'ALL' collection, to identify the most used words of all
+        multi.sadd('____all____', s); // add keyword
+        multi.incrby('____all____'+':'+s, owner); // track its density
+    }
+    else {
+        client.sadd('____all____', s); // add keyword
+        client.incrby('<'+':'+s, owner); // track its density
+    }
+};
+
+Db.prototype.getTopRelations = function(owner, callback, res) {
+    client.sort('____all____', "by", "____all____:*", 'LIMIT', 0, 500, 'DESC', "get", "#", function (err1, items1) {
+        // get most often realted keywords for the given keyword
+        client.sort(owner, "by", owner+":*", 'LIMIT', 0, 120, 'DESC', "get", "#", function (err2, items2) {
+          // remove the noise by removing the most often used keywords
+          callback(tools.inAButNotInB(items2,items1),res);
+        });
+    });
+};
+
+Db.prototype.flush = function(callback) {
+    if(false === isMulti || multi === null) {
+        console.log('Error - you must use enableMulti() before flush().');
+        return;
+    }
+    callback = callback || null;
+    multi.exec(callback);
+    // that.disableMulti();
+    multi = null;
+    isMulti = false;
+};
+
+module.exports = Db;
\ No newline at end of file
diff --git a/app/http.js b/app/http.js
new file mode 100644
index 0000000..5ff51b0
--- /dev/null
+++ b/app/http.js
@@ -0,0 +1,50 @@
+var restify = require('restify');
+var express = require('express');
+var Db = require('./db');
+var config = require('../config');
+
+var Http = function() {
+    server = restify.createServer();
+    server.use(restify.bodyParser());
+    db = new Db();
+    that = this;
+
+    // initialise routes
+    that.initRoutes();
+    that.init();
+};
+
+Http.prototype.initRoutes = function() {
+    if(config.creds.debug) {
+        console.log('initRoutes');
+    }
+    // Set up our routes
+    server.get('/relations/:name', that.responseRelations);
+};
+
+Http.prototype.init = function() {
+    if(config.creds.debug) {
+        console.log('start server');
+    }
+    // start the server
+    server.listen(config.creds.server_port, function() {
+      console.log('%s listening at %s', server.name, server.url);
+    });
+};
+
+Http.prototype.responseRelations = function(req, res, next) {
+    if(config.creds.debug) {
+        console.log('responseRelations for "'+req.params.name+'"');
+    }
+    console.log('Will deliver top relations for requested word "'+req.params.name+'".');
+    db.getTopRelations(req.params.name, that.doResponse, res);
+};
+
+Http.prototype.doResponse = function(data, res) {
+    if(config.creds.debug) {
+        console.log('doResponse');
+    }
+    res.send(data);
+};
+
+module.exports = Http;
\ No newline at end of file
diff --git a/app/model.js b/app/model.js
new file mode 100644
index 0000000..e957424
--- /dev/null
+++ b/app/model.js
@@ -0,0 +1,13 @@
+var Model = function(s) { 
+  val = s || null;
+};
+
+Model.prototype.getValue = function() { 
+  return val; 
+};
+
+Model.prototype.setValue = function(s) { 
+  val = s; 
+};
+
+module.exports = Model;
\ No newline at end of file
diff --git a/app/scraping.js b/app/scraping.js
new file mode 100644
index 0000000..772ef97
--- /dev/null
+++ b/app/scraping.js
@@ -0,0 +1,149 @@
+var Db = require('./db');
+var Analyse = require('./analyse');
+var restify = require('restify');
+var config = require('../config.js');
+var $ = require("jquery");
+
+var Scraper = function() {
+    db = new Db();
+    analyse = new Analyse();
+    that = this;
+    debug_once = false;
+    wikipedia = restify.createJsonClient({
+      url: 'http://'+config.creds.lang+'.wikipedia.org',
+      version: '*'
+    });
+};
+
+Scraper.prototype.wikiSearch = function(s) {
+    if(config.creds.debug) {
+        console.log('wikiSearch');
+    }
+
+    // check if not empty string
+    if(typeof s == 'undefined' || s === '' || debug_once) {
+        console.log('wikiSearch - the given string is empty or undefined!');
+        console.log('node will exit now.');
+        process.exit(1);
+        return;
+    }
+
+    wikipedia.get('/w/api.php?action=opensearch&search='+escape(s)+'&format=json&limit=3', function(err, req, res, data) {
+        if(typeof data[1] == 'undefined' || typeof data[1][0] == 'undefined') {
+          if(config.creds.debug) {
+            console.log('No page found in wikipedia for '+req.path);
+          }
+          db.removePageFromQueue(s);
+          that.goToNext();
+          return;
+        }
+
+        // get first matching result
+        var firstTitle = data[1][0];
+
+        // set first result as done
+        db.addPageAsDone(firstTitle,function (err, result) {
+            if(config.creds.debug) {
+                console.log('addPageAsDone callback');
+            }
+            if(result) {
+                that.wikiGrab(firstTitle);
+                db.removePageFromQueue(firstTitle);
+            }
+            else {
+                if(config.creds.debug) {
+                    console.log(firstTitle+' was crawled already!');
+                }
+                that.goToNext();
+                return false;
+            }
+        });
+
+        // add all sites to queue
+        for (var i = data[1].length - 1; i >= 0; i--) {
+          console.log('Added '+data[1][i]+' to queue!');
+          db.addPageToQueue(data[1][i]);
+        }
+
+        if(config.creds.debug) {
+            debug_once = true;
+        }
+    });
+};
+
+Scraper.prototype.wikiGrab = function(s) { 
+    if(config.creds.debug) {
+        console.log('wikiGrab with '+s);
+    }
+    wikipedia.get('/w/api.php?rvprop=content&format=json&prop=revisions|categories&rvprop=content&action=query&titles='+escape(s), function(err, req, res, data) {
+        if(typeof data.query == 'undefined') {
+          that.goToNext();
+          return false;
+        }
+
+        // check if valid content
+        if(typeof data.query.pages[Object.keys(data.query.pages)[0]].revisions == 'undefined') {
+          that.goToNext();
+          return false;
+        }
+
+        // get the main content of the wikipedia page
+        var rawtext = data.query.pages[Object.keys(data.query.pages)[0]].revisions[0]["*"];
+         // now split the whole content into text blocks
+        var parts = rawtext.split(/\n|\r/);
+        var snippets = [];
+
+        if(config.creds.debug) {
+            console.log('going to http://'+config.creds.lang+'.wikipedia.org/wiki/'+s);
+        }
+
+        // loop all text blocks and pull these with more than config.creds.min_text_block_length (default: 120) chars
+        for (var i = parts.length - 1; i >= 0; i--) {
+          if(parts[i].length > config.creds.min_text_block_length) {
+            snippets.push(parts[i]);
+          }
+        }
+
+        console.log('Will now process "'+s+'" with '+snippets.length+' text blocks.');
+
+        if(snippets.length > 0) {
+          // give the loop worker something to do
+          that.loopWorker(snippets);
+        }
+        else {
+          // restart fetch
+          that.goToNext();
+        }
+
+    });
+};
+
+Scraper.prototype.loopWorker = function(snippets) { 
+    if(config.creds.debug) {
+        console.log('loopWorker');
+    }
+    // when snippetbox is empty, restart fetch
+    if(snippets.length === 0) {
+        if(config.creds.debug) {
+            console.log('Finished all Snippets!');
+        }
+        that.goToNext();
+        return;
+    }
+
+    // analyze full text block
+    $.when(analyse.scanTextBlock(snippets.pop(),snippets.length)).done(function() {
+        // set a timeout to be gently to the memory and cpu 
+        // (can be changed in the config file)
+        var t=setTimeout(function(){that.loopWorker(snippets);},config.creds.sleeptime);
+    });
+};
+
+Scraper.prototype.goToNext = function() { 
+    if(config.creds.debug) {
+        console.log('goToNext');
+    }
+    db.getRandomItemFromQueue(that.wikiSearch);
+};
+
+module.exports = Scraper;
\ No newline at end of file
diff --git a/classes/apiprovider.js b/classes/apiprovider.js
deleted file mode 100644
index 7826482..0000000
--- a/classes/apiprovider.js
+++ /dev/null
@@ -1,88 +0,0 @@
-module.exports = function(client) {
-  // var that = this;
-
-  return {
-    /** 
-     * function inAButNotInB will remove all items from array a which are in array b
-     * depending on underscore.js
-     *
-     * @param
-     * @return
-     */
-    inAButNotInB: function (A, B) {
-      return _.filter(A, function (d) {
-        return !_.contains(B, d);
-      });
-    },
-
-    Base: Base,
-
-    /** 
-     * function doResponse will send the response to the client
-     *
-     * @param string data
-     * @return res
-     */
-    doResponse: function (data, res) {
-      res.send(data);
-    },
-
-    /** 
-     * function getRelations will take action as a router function to deliver all relations to the requested keyword
-     *
-     * @param string req.params.name
-     * @return boolean
-     */
-    getRelations: function (req, res, next) {
-      var base = new Base(req.params.name,client);
-      base.setRes(res);
-      base.getTopRelations();
-    }
-  }
-};
-
-/** 
- * class Base will get handle database-actions related to one keyword
- *
- * @param string val
- * @return boolean
- */
-Base = function (val,client) {
-  // Store variables
-  var that = this,
-      multi_in = client.multi(), // to pipeline actions for redis
-      res;
-
-  // to set the restify response, a bit hacky actually
-  this.setRes = function(val) {
-    res = val;
-  };
-
-  // process the pipelined actions in redis
-  this.save = function() {
-    multi_in.exec();
-  };
-
-  // get all relationes, without the noise
-  this.getTopRelations = function() {
-    // get most often used keywords (limit 500)
-    client.smembers('____all____', function (err1, items1) {
-      console.log(items1);
-      return;
-        // get most often realted keywords for the given keyword
-        client.sort(val, "by", val+":*", 'LIMIT', 0, 120, 'DESC', "get", "#", function (err2, items2) {
-          // remove the noise by removing the most often used keywords
-          doResponse(inAButNotInB(items2,items1),res);
-        });
-    });
-  };
-
-  // add word and count up
-  this.pushRelation = function(rel, incr) {
-    multi_in.sadd(val, rel);
-    if(typeof incr == 'undefined') {
-      incr = 1;
-    }
-    multi_in.incrby(val+':'+rel, incr);
-  };
-};
\ No newline at end of file
diff --git a/classes/textanalyze.js b/classes/textanalyze.js
deleted file mode 100644
index 8d7625e..0000000
--- a/classes/textanalyze.js
+++ /dev/null
@@ -1,70 +0,0 @@
-module.exports = function(client) {
-
-  return {
-    /** 
-     * function analyzeText will get the content for the given wikipedia page title
-     *
-     * @param string title
-     * @return boolean
-     */
-    analyzeText: function (snippet,counter) {
-      // split the text block to words
-      var words = tools.tokenize(snippet);
-
-      if(config.creds.debug)
-          console.log('Count of words in snippet ('+counter+'): '+words.length);
-
-      // create empty object
-      var obj = {};
-
-      var multi = client.multi();
-
-      // loop all words
-      for (var i = words.length - 1; i >= 0; i--) {
-
-        // count all seen words
-        if(typeof obj[words[i].toLowerCase()] == 'undefined')
-          obj[words[i].toLowerCase()] = 1;
-        else
-          obj[words[i].toLowerCase()]++;
-
-        // add every word to the queue to spread the scrape
-        multi.sadd('____sites2do____',words[i].toLowerCase());
-
-        // if(config.creds.debug)
-        //   console.log(words[i].toLowerCase()+'¥ - '+words[j].toLowerCase()+' - '+similar_text(words[i].toLowerCase(),words[j].toLowerCase(),1));
-      }
-
-      var base;
-
-      $.each(obj, function(index, value) {
-
-        // skip if not valid
-        if(typeof index == 'undefined' || typeof index.toLowerCase == 'undefined')
-          return;
-
-        // create new obj from class Base, make sure to work with lowercase only
-        base = new Base(index.toLowerCase());
-
-        // loop all words
-        $.each(obj, function(index2, value2) {
-          if(index != index2) {
-            // add relation, value2 is the counter of how often the word was seen in the recent textblock
-            base.pushRelation(index2.toLowerCase(),value2);
-          }
-        });
-
-        base.save();
-
-        // add to our general 'ALL' collection, to identify the most used words of all
-        multi.sadd('____all____', index.toLowerCase()); // add keyword
-        multi.incrby('____all____'+':'+index.toLowerCase(), value); // track its density
-
-      });
-
-      multi.exec(function(err, replies) {
-          return true;
-      });
-    }
-  }
-};
\ No newline at end of file
diff --git a/classes/wiki.js b/classes/wiki.js
deleted file mode 100644
index 7581d77..0000000
--- a/classes/wiki.js
+++ /dev/null
@@ -1,132 +0,0 @@
-module.exports = function(client) {
-  return {
-    /** 
-     * function wikiSearch will start the main processes to search for the best wikipedia page for the given string
-     *
-     * @param string term
-     * @return boolean
-     */
-    wikiSearch: function (term) {
-      // do api call
-      wikipedia.get('/w/api.php?action=opensearch&search='+escape(term)+'&format=json&limit=3', function(err, req, res, data) {
-
-        if(typeof data[1] == 'undefined' || typeof data[1][0] == 'undefined') {
-          if(config.creds.debug)
-            console.log('No page found in wikipedia for '+req.path);
-          client.srem('____sites2do____',term);
-          goToNext();
-          return;
-        }
-
-        // get first matching result
-        var firstTitle = data[1][0];
-
-        // set first result as done
-        client.sadd('____sites____', firstTitle, function (err, result) {
-          if(result) {
-            wikiGrab(firstTitle);
-            client.srem('____sites2do____',firstTitle);
-          }
-          else {
-            if(config.creds.debug)
-              console.log(firstTitle+' was crawled already!');
-            goToNext();
-            return false;
-          }
-        });
-
-        // add all sites to queue
-        for (var i = data[1].length - 1; i >= 0; i--) {
-          client.sadd('____sites2do____',data[1][i]);
-        }
-
-      });
-    },
-
-    /** 
-     * function wikiGrab will get the content for the given wikipedia page title
-     *
-     * @param string title
-     * @return boolean
-     */
-    wikiGrab: function (title) {
-      // do the api call
-      wikipedia.get('/w/api.php?rvprop=content&format=json&prop=revisions|categories&rvprop=content&action=query&titles='+escape(title), function(err, req, res, data) {
-        if(typeof data.query == 'undefined') {
-          goToNext();
-          return false;
-        }
-
-        // check if valid content
-        if(typeof data.query.pages[Object.keys(data.query.pages)[0]].revisions == 'undefined') {
-          goToNext();
-          return false;
-        }
-
-        // get the main content of the wikipedia page
-        var rawtext = data.query.pages[Object.keys(data.query.pages)[0]].revisions[0]["*"];
-         // now split the whole content into text blocks
-        var parts = rawtext.split(/\n|\r/);
-        var snippets = [];
-
-        if(config.creds.debug)
-          console.log('going to http://'+config.creds.lang+'.wikipedia.org/wiki/'+title);
-
-        // loop all text blocks and pull these with more than config.creds.min_text_block_length (default: 120) chars
-        for (var i = parts.length - 1; i >= 0; i--) {
-          if(parts[i].length > config.creds.min_text_block_length) {
-            snippets.push(parts[i]);
-          }
-        }
-
-        if(snippets.length > 0) {
-          // give the loop worker something to do
-          loopWorker(snippets);
-        }
-        else {
-          // restart fetch
-          goToNext();
-        }
-
-      });
-    },
-
-    /** 
-     * function loopWorker will process all snippets gently for your system
-     *
-     * @param array snippets
-     * @return
-     */
-    loopWorker: function (snippets) {
-      // when snippetbox is empty, restart fetch
-      if(snippets.length === 0) {
-        if(config.creds.debug)
-          console.log('Count of snippets: '+snippets.length);
-        goToNext();
-        return;
-      }
-
-      // analyze full text block
-      $.when(text.analyzeText(snippets.pop(),snippets.length)).done(function() {
-        // set a timeout to be gently to the memory and cpu 
-        // (can be changed in the config file)
-        var t=setTimeout(function(){loopWorker(snippets);},config.creds.sleeptime);
-      });
-    },
-
-    /** 
-     * function goToNext will move on to a random element to search for in the queue ____sites2do____ which is stored in redis
-     *
-     * @param
-     * @return
-     */
-    goToNext: function () {
-      if(config.creds.debug)
-        console.log('NEXT');
-      client.smembers('____sites2do____', function (err, result) {
-        var randomnr=Math.floor(Math.random()*result.length);
-        wikiSearch(result[randomnr]);
-      });
-    }
-  }
-};
\ No newline at end of file
diff --git a/config.js b/config.js
index db832b8..ed9e38a 100644
--- a/config.js
+++ b/config.js
@@ -1,11 +1,12 @@
 // Don't commit this file to your public repos
 exports.creds = {
+  http_server: false, // default: false
   server_port: 8080, // default: 8080
   sleeptime: 0, // sleep time in ms between the iteration of all textblocks
   mongoose_auth_local: 'mongodb://localhost/opensemanticapi', // default: mongodb://localhost/opensemanticapi
   min_text_block_length: 120, // default: 120
-  debug: true, // default: true
+  debug: false, // default: true
   lang: 'en' // default: english
   // lang: 'de' // german
-  // lang: 'es' // spansih
+  // lang: 'es' // spanish
 };
\ No newline at end of file
diff --git a/lib/tools.js b/lib/tools.js
index d111e8f..4012db4 100644
--- a/lib/tools.js
+++ b/lib/tools.js
@@ -1,5 +1,11 @@
+var _ = require("underscore");
 
 module.exports = { 
+  inAButNotInB: function (A, B) {
+    return _.filter(A, function (d) {
+      return !_.contains(B, d);
+    });
+  },
   similar_text: function (first, second, percent) {
     // http://kevin.vanzonneveld.net
   // +   original by: Rafał Kukawski (http://blog.kukawski.pl)