Dictation software for Windows · GDPR compliant Dictate, read documents, take meeting notes. 3× faster. 100% local.

Download SpeechToWork using the button above and install it, no administrator rights needed. On first start, choose “Try free for 14 days”, enter your name and business email address and off you go. You are not asked for payment details and there is nothing to cancel. If you already have a licence key, enter it in the same place. On first start, SpeechToWork downloads the language models once, 4 to 6 GB depending on your computer; after that a short introduction walks you through the app.

The trial ends automatically. Without a licence you can no longer use SpeechToWork, but nothing is charged and nothing renews. Your settings, your dictionary and all the files you created stay on your computer. As soon as you take out a subscription and enter your licence key, you carry on exactly where you left off.

SpeechToWork processes speech and text only on your own computer. Nothing is uploaded, neither to us nor to a cloud provider. That means no third party processes your dictations and there are no international data transfers. In your record of processing activities, it is listed simply as locally installed software. This applies under both the UK GDPR and the EU GDPR.

Only for two things: the one-off download of the language models on first start and checking your licence. After that, dictation, writing, documents and audio, meeting minutes and reading receipts all work without the internet, on the train or on an isolated network.

Any program with a text field: Outlook, Word, Excel, Teams, your browser, your CRM, your practice or case management software. Hold a key, speak, let go, and the text appears at the cursor. No integration or plug-in needed.

A current PC or laptop with Windows 10 or 11; 16 GB of RAM is recommended. On start, SpeechToWork checks the graphics card and memory and picks the right language model itself. With a dedicated graphics card of 6 GB or more, the finished text appears after about a second; on computers without one, after a few seconds.

Classic speech to text software writes down what you say. SpeechToWork also writes it up properly: filler words disappear, self-corrections are applied, numbers and punctuation are right. Unlike cloud services, it sends nothing to the internet to do so. It needs no voice training and costs a fixed price per user instead of an expensive single licence.

With dictation, you speak the text you want to end up on the page, and SpeechToWork polishes it. A call note is for the moment right after a phone call: tap the record key twice, choose “Note” and talk freely. You get a structured note with the contact, topic, key facts and next steps. It is saved, and your own tasks go on your task list.

Yes. Select an email or a paragraph, hold the record key and say “read this to me”. SpeechToWork can also read out answers to your questions and document summaries if you like. The voice is generated on your computer, nothing is uploaded. There are three voices to choose from, and you can switch speech output off at any time.

Yes. Drag in a PDF, Word document or text file, or tap the record key twice and choose “Document”. You get a summary in a fixed structure: what it is about, the key points, deadlines and dates, amounts, what to do and what to watch out for. Every point comes with its source, such as [Page 2]. Deadlines like “three months before the end of the term” are converted into an exact date by the software. Then you can ask questions, typed or spoken. If the answer is not in the document, SpeechToWork says so instead of making something up.

No. Contracts, official letters, voicemails, interviews and videos are processed entirely on your computer. There is no upload, neither to us nor to an AI provider. The transcript and summary of a recording are also saved as a text file on your hard drive. With a graphics card, a 20-page document is summarised in about 10 to 20 seconds, and answers to questions take about a second.

Only with the consent of everyone taking part. The rules differ from country to country, and in Germany, for example, prior consent is a legal requirement. SpeechToWork reminds you before every recording. Because the recording, transcript and minutes never leave your computer, there is no additional service provider you need to tell participants about.

Yes. In the settings, click the field for the function and press the key you want, SpeechToWork applies it straight away. Keys you don't need for typing work best, such as function keys, Right Ctrl, Pause or the number pad keys. You choose the microphone from a list, for example a headset, conference microphone or webcam, and check it with a level test.

No. The price per user is all you pay. The AI runs on your own computer, not on a provider's servers, so there are no fees per minute, per word or per request, and no second subscription to an AI service. You can dictate as much as you like. Only VAT is added to the net price.

25 European languages, including English, German, French, Spanish, Italian, Polish and Dutch. The language is detected automatically. Say “now all of that in French” at the end of a dictation and you get the text translated straight away. The app itself is available in English and German.

SpeechToWork is currently available for Windows 10 and 11, because that is where most office workstations run. A macOS version is planned. Send a short message with the keyword Mac to kontakt@speechtowork.de and we will let you know when it launches.