Slang TTS

Slang TTS is the actual core of our technology, the text-to-speech engine. It can be applied in client/server solutions as well as a static component in an embedded system or in a larger application. Languages and dialects (the "lingware") can be loaded as needed, without having to exchange the TTS engine.

WebSlang: Rapid Prototyping

In order to use WebSlang, you simply send a text to our server and you will get back a corresponding wav file. This service is available free of charge for the time being and is provided as an instrument for your application development.

The use of WebSlang in your own application is very easy:

View result in browser
<!DOCTYPE html>
    <title>SlowSoft TTS Example</title>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1">

    <script src=""></script>
      var tts = new TTS();
      sendTtsRequest = function () {
        text = document.getElementById('text').value
        lang = document.getElementById('language').value
        tts.toAudio(text, lang);
      <form id="inputform" name="TTS" method="post">
        <select id="language" name="language">
          <option value="gsw-CHE-ba"> Baseldeutsch   </option>
          <option value="gsw-CHE-be"> Berndeutsch    </option>
          <option value="gsw-CHE-gr"> Bündnerdeutsch </option>
          <option value="gsw-CHE-lu"> Luzerndeutsch  </option>
          <option value="gsw-CHE-sg"> St. Gallerdeutsch  </option>
          <option value="gsw-CHE-zh"> Zürichdeutsch  </option>
          <option value="roh-CHE-vl"> Vallader       </option>
          <option value="deu-CHE">    Schriftdeutsch </option>
          <option value="eng-USA">    US-English     </option>
        <textarea id="text" name="text"></textarea>
      <button onclick="sendTtsRequest();">OK</button>

Please note that our server has only limited capacities and may be offline during short intervals of maintenance. Moreover, it cannot be guaranteed that the same texts always yield identical sound files, since we will continuously improve the quality of the speech output. WebSlang is therefore mainly intended as a tool for rapid testing and prototyping.

SiteSlang: TTS on your own servers

You install the Slang TTS engine on your servers, together with a client/server interface provoided by SlowSoft (in preparation).

SlangDK: Full Integration of TTS into Your Application

The Slang TTS engine is fully integrated into your own (embedded) application via an API and in the form of a static or dynamic library (availability upon request).

Lingware (language and speech data)

For all Slang TTS products, the following lingware modules are currently available:

  • Swiss German dialect (Grisons dialect, female voice Selina)
  • Rhaeto-Romanic (Vallader idiom, female voice Seraina)
  • Swiss Standard German (female voice Selma)
  • Swiss German dialect (Zurich dialect, male voice Zeno)
  • US English (female voice Amanda)

The first three modules were created from the same voice and can thus be tightly coupled and mixed.

Also, the following lingware modules are available in preview versions:

  • Swiss German dialect (Basel dialect, female voice Barbara)
  • Swiss German dialect (Bern dialect, male voice Beat)
  • Swiss German dialect (Lucerne dialect, male voice Lukas)
  • Swiss German dialect (St Gall dialect, female voice Sarah)

For all of the voices there are also more "robotic" sounding variants which have been derived from the original voices. They may be better suited for certain robotics applications than the original voices.

Input Converters for Swiss German

Due to the lack of a writing standard for Swiss German we offer different possibilities for Swiss German text entry:

  1. Pseudo-standard German input

    This format is based on Standard German and is especially well suited for the creation of dialog systems. In this format the Swiss German text is represented word by word by corresponding Standard German (or in few cases artificial) words. In the case of ambiguities it is possible to specify more precisely some grammatical features by markups, e.g., word category, case, number, gender, and/or person of a word.

    Example 1:

    "Bitte gib dem Peter die Nummer von deiner Bestellung."

    (vs. Standard German "Bitte gib Peter die Nummer deiner Bestellung.")

    Example 2:

    "Bitte lug da auf die Strasse."

    (vs. Standard German "Bitte schau hier auf die Strasse.")

    Example 3:

    "gehen#p3"  (3rd plural form desired, not infinitive)

    The advantage of this format compared to a phonetics-based Swiss German input format is (to a certain degree) its dialect independence. That is, the conversion from the Standard German words to Swiss German dialect words is done automatically according to a desired dialect.

  2. Standard German input

    This format is currently under development and is used in the demo page. In this format, Standard German text is translated into Swiss German as much as possible. Thus, for example, past tense is automatically replaced by perfect tense (e.g., "suchte" → "habe ... gesucht"), person names are preceded by articles (e.g., "Peter" → "der Peter"), and Standard German words that do not exist in Swiss German are replaced by Swiss German substitutes (e.g., "arbeiten" → "schaffen").

    Example: "Peter arbeitete gestern bis in die Nacht." is converted to "Der Peter hat gestern bis in die Nacht geschafft."

  3. Swiss German input

    This format will be developed in the near future. In this format a dialect-specific, non-standard pseudo-phonetic writing will be made available, as it is used in SMS and other messaging systems.

    Example: "ir kömend aber spoot" (Chur dialect) or "dir chömid aber spät" (Bern dialect)


In order to tailor the Slang TTS system according to your needs and thereby to optimize the speech quality, we offer the following services:

  • lexicon optimization (e.g., pronunciation of application-specific person and place names)
  • optimization of frequent application-specific phrases in our language and speech data for highest output quality
  • establishing of new languages and dialects (with focus on Swiss needs)
  • creation of new voices

Integration/implementation of full applications

We can offer the integration of our TTS system into overall applications, i.e., together with speech recognition and a dialog component, via a partner company. For speech recognition in Swiss German and other languages we collaborate with the company Spitch AG.