We are here to help you

An overview you can find here:

https://www.gospeech.com/en/prices

The GoSpeech free plan can be used indefinitely. However, it is necessary to log in to GoSpeech regularly to avoid account inactivity.

For sustainability reasons, GoSpeech reserves the right to delete accounts that have not been logged in for 3 months. Prior to deletion, users will receive 3 reminder messages: 1 month, 1 week and 1 day before the account is scheduled to be deleted. After deletion, the account and data cannot be restored. This approach serves to reduce data waste and contributes to the sustainable use of resources.

The free plan is suitable for testing the GoSpeech transcription software for the first time.

With it, you can upload three files per month with a maximum length of 10 minutes and, in addition to the excellent quality of speech recognition, you can also take advantage of the many useful editing functions of our online editor. We wish you a lot of fun while testing!

Your files created in the Free Plan will be retained unless you have deleted your account or your account has been automatically deleted due to 3 months of inactivity. If you sign up for a new month or switch to another plan, your files will still be available.

You can easily upload files longer than 10 minutes even in the Free plan. GoSpeech automatically truncates the audio or video file after 10 minutes, so only the first 10 minutes are transcribed. So you don't have to worry about the length of your files. If you want to transcribe a file longer than 10 minutes, you can simply switch to the prepaid plan. Buy as much time as you need here.

Your files will remain in your account.

*If you switch from a standard plan to a free account, we reserve the right to delete the account for sustainability reasons if you have not logged in for more than 3 months. We will of course inform you in good time before deletion to give you the opportunity to prevent this.

You can switch to a higher plan at any time.

You can switch to a lower plan taking into account the minimum duration of your existing plan.

The minimum term depends on your chosen plan. Please see our table for more information:

The inclusive hours in the monthly subscription can be used in the current month.

Unused hours will not be transferred to the following month.

Log in to your GoSpeech account. In the account settings under "current plan", you can cancel the subscription respecting the minimum term.

Yes, you can purchase additional hours with any pricing model. In the standard subscription model, you can redeem the purchased additional hours until the end of the subscription month. With a prepaid booking, the additional hours remain available for up to 3 years.

Yes, as a company or organization, you can try GoSpeech. For future Enterprise customers, we offer the option to request a trial quota. To do so, please fill out the form.

 Paypal, direct debit, credit card

You have a period of 3 years to use up the prepaid hours.

No, absolutely not.

User data and data from audio and text files stored in Go Speech are not shared with other companies.

The data is stored in a German data center DSGVO compliant and not evaluated.

When you upload a file, the different statuses of your file change depending on the process step. First, the file is uploaded - status: "Uploading".  

After successful upload, the file is sent for speech recognition and the status is: "In recognition". 

If the file is speech recognized and ready to be edited and corrected by you, then the status changes to "Edit".  

As soon as you start editing the file, the status changes to "In progress". 

When editing is finished, you can set the status of the file to "Done" yourself by clicking "Complete editing" in the editor. 

Audio and video files are converted to text in the order in which they are uploaded. Currently, manual prioritization is not possible.

The duration of the transcription depends on the length of the audio recording.

As a rule of thumb, the formula Duration of transcription in minutes = Audio length in minutes / 2 can be used as a rule of thumb.

A low internet upload capacity prolongs the upload and thus also the time it takes until the recognised text is available.

No. However, the speech recognition results can be used to improve the system. This is done without content evaluation and without reference to the user.

If the Internet connection is interrupted while uploading an audio or video file, the upload process stops. Once the connection is restored, the upload process continues. 

If the web page is updated during the upload process, a message appears indicating the further procedure. In case of an update, the upload process is aborted, already uploaded file parts are not saved.

GoSpeech creates backups of the transcript after changes are made to the transcript. The version name is displayed as the date and time of the save. These can be viewed and restored in the version history. 

To restore, click the "Save as Copy" button next to the version number.

The recovery can also be called from the editor.

A copy of the version is created in the recording list, which can now be edited normally. The copy has the designation "Recovered copy of <title> at <name of version>".

The new recording can now be edited as usual.

GoSpeech creates backups of subsequent changes to the transcript, which are displayed in the version history: 

• Speaker change 
• Formatting change 
• Text changes

In the GoSpeech Editor, keywords are determined and output with every audio and video recording. By clicking on the "W" in the upper right corner of the "Search" field, you can access the keywords. If you click on one of the keywords, this word will be displayed with a green border in the text.

If you want to add a keyword yourself, you have two options. In both cases, the added word will subsequently appear in the list of keywords. On the one hand, you can select the desired word in the transcript and add it using the plus button that appears.

On the other hand, you can use the pencil button in the upper right corner next to the keywords to open a menu that allows you to search for specific keywords or add new ones. To add a new keyword, enter it in the input field and then click on Create. Within the menu you also have the possibility to edit or delete the keywords you have found or added. To do so, simply click on the respective word in the list. 

The recognised text is edited in the GoSpeech online editor. Audio and text are kept synchronised in the transcript. This allows quick highlighting, editing and correcting or searching of your transcript.

The Notpad is the practical and clear extension of the Online Editor.

You can easily drag text parts or sections from the transcript into the Notpad and compile them into a new document - exactly according to your needs. For example, you can quote text passages from a transcribed interview and incorporate them into an article in Notepad and then export the article. The article and the complete version of the interview remain intact.

The GoSpeech Editor contains a transcript and a notepad. When the recognized text from the recording list is opened for the first time, only the transcript is displayed. Icons are available in the icon bar to show and hide the transcript and notepad.

Recognized (and possibly already corrected) text can be exported by clicking the export icon. The export can be started from the work list as well as from the editor.

After clicking, a window opens in which the file format can be selected.

Furthermore, it can be selected whether time stamps and/or speaker names are to be exported as well.

If speaker changes are not recognised automatically, they can easily be added later. To do this, click in front of the first word of the new speaker and press the key combination "ALT + ENTER". The following text will now be assigned to a new speaker.

Similarly, you can also merge speaker segments of two speakers and assign them to the first of the two speakers.

This might be necessary if the system has detected a speaker change where no change has taken place.

To do this, simply click in front of the first word of the incorrectly recognised speaker and operate the delete key.

The incorrectly detected speaker change is thus deleted, the text of both speakers is merged and assigned to the first speaker.

The key combination ALT+ENTER inserts a new segment with a speaker in the editor.

The new speaker is positioned after the cursor.

Vocabulary can be used to increase the probability that a spoken word will be assigned the desired individual spelling. This is particularly useful for specialized vocabulary, proper names, and words with different spellings. Created vocabularies have one written and up to ten spoken forms.

A vocabulary word has exactly one written form. This is the written form that should be set automatically in the transcript when one of the up to ten stored spoken forms for this vocabulary is said in the audio or video track.

A brief example for better understanding:

Written form: gnocchi

Spoken forms: nyohki; nokey; nochi; gnotschi

No, but spoken forms significantly increase the likelihood that a word will be spelled as desired.

Yes, the written form of a vocabulary item can be created multiple times. Spoken forms, on the other hand, are unique and therefore can only be added to a written form.

Up to 1000 vocabularies can be created per language.

A vocabulary can be created either for a specific language or for all languages. This makes sense for proper names, for example. If a vocabulary is to be created for several, but not all languages, the vocabulary must be created for each language individually.

In GoSpeech, individual vocabulary items are managed in dictionaries in the sense of a folder structure. This allows you to flexibly decide which dictionaries should be used for the transcription of a file.

Up to ten dictionaries per user are possible.

In the upload mask under "further settings" , "vocabulary" can be selected which dictionaries should be used for the transcription of the uploaded file. It is not possible to apply the vocabulary after the transcription.

Therefore, please be sure to check in advance which dictionaries are to be used for the respective file.

An applied vocabulary does not provide a hundred percent guarantee that a word will be spelled correctly, as the pronunciation in the audio track may differ slightly. However, the probability for the given spelling is significantly increased. To further influence this probability, additional spoken forms can be added to a word.

Feel free to use our feedback function by highlighting the word in the transcript, clicking on the blue plus symbol and selecting "Suggest another spelling". In the mask that opens, you can give us a suggestion for the correct spelling of the word. We will check the case promptly and adjust our speech recognition so that the word will be recognized correctly in the future.

There are three possibilities. You can create vocabularies manually one by one by selecting a dictionary in the vocabulary menu and creating a new entry, or you can select a word in a transcript, click on the blue plus symbol and select "add to vocabulary" and create a new entry here. If several vocabularies are to be added to a dictionary, it is recommended to import vocabulary lists.

Import of a new dictionary or import of several vocabularies into an existing dictionary is possible.

The correct file format is essential. Currently it is possible to import xls, xlsx, or separated csv.

The first line contains the headings of the columns:

"word": Exactly one written form for the word must be stored here.

"spoken form": Up to ten different forms can be added here. Please separate individual forms with a semicolon.

"language": The language of the word must be entered here.

Please use the country code list for this (ISO-3166-1 code list - Wikipedia). If all languages are to be set, "all" must be used.

Here is an example of a valid import table:

Currently it is possible to import .xsl, .xslx and .csv files.

You have the option to skip the words where an error message occurs and perform the import without them. Or you can cancel the import, fix the errors and run the import again.

Your employee can export their dictionaries and provide them with the file. They can then import this file to their account.

Currently, files with the following formats can be processed for recognition:

• aac
• flac
• m4a
• m4v
• mp3
• mp4
• ogg
• wav
• dss
• ds2

GoSpeech currently supports speech recognition in the following languages: 

• German 
• English (Global) 
• British English 
• US English 
• Australian English 
• French 
• Danish 
• Swedish 
• Norwegian 
• Finnish
• Italian 
• Spanish 
• Portuguese 
• Russian 
• Ukrainian 
• Turkish 
• Dutch 
• Polish 
• Czech 

Yes. When you upload a recording, you can select the main language of the recording.

You have the possibility to export subtitles either in .vtt format or embed them as a part of the video in video format.

The following video formats are supported: .mp4, .mv4a, .ogg.

Video export is available only for users of the Prepaid, Standard or Enterprise plan.

The creation of the video may take some time depending on the length of the original video. We will notify you by email and send you a download link once the video is ready.

The following video shows the whole video export process: Link

It can happen that emails unintentionally end up in the spam folder or are blocked by the firewall, which prevents them from showing up in your inbox.

Here are a few tips to prevent this from happening:

Add senders to your address book

All emails sent by GoSpeech have the following domain @gospeech.com. You might therefore consider adding this domain to your address book or ask your network administrator to do so.

Check SPAM folder

Emails from GoSpeech can get stuck in SPAM filters. Check the SPAM folder in your email inbox to see if the message ended up there. If it is not in your local SPAM folder, ask your network administrator if the message may be blocked at the network level.

Consult with your administrator

If you have tried the above recommendations and still cannot receive mail from GoSpeech, you may need to contact your administrator to see if your domain is reachable.

If you are not receiving mail despite this measure, please inform us about it via support@gospeech.com.

Multi-channel recordings are recordings that simultaneously record several audio signals from different sources or channels. Typically, these are different microphones.

Such recordings are often made in large meeting rooms or during music or film productions.

To create multi-channel recordings, you need special hardware components. A simple example: microphones or other audio sources send sound information via cable to an audio interface that can process multiple channels. From there, the data is transferred to the recording software on the computer.

Make sure that the individual audio tracks do not record the conversations of other participants, as this can impair the recognition results.

The following graphic illustrates the recognition result when the microphones record other participants in the room: 

Synchronized recording:

Make sure that all microphones are synchronized to ensure that the recorded audio data is timed. This is crucial so that the data from the different channels can be combined correctly.

One potential error is that individual speakers are recognized multiple times, which leads to inaccurate transcription (see diagram).

This can be avoided by ensuring that each microphone only records the intended speaker and that no secondary speakers are recorded.

The GoSpeech API is a speech-to-text interface that allows you to integrate advanced speech recognition technology into your software or existing systems.  

You can request access to a free trial environment here to try out the API or to send us a no-obligation enquiry for integration into your system directly.  

Yes, in addition to the cloud option, we also offer an API function for our on-premises solutions. The API functionality is already included in the price.

Yes, it runs in a GDPR-compliant environment on German servers and offers an on-premises option.

All languages supported by GoSpeech are also available in the API. You can find more information under FAQ.

When using the cloud version, you always have access to the latest software version without additional update fees.

You can find an introduction to using the API in our API documentation. 

The API documentation is available here on request.

Our API does not currently have the option of multichannel transcription. Please contact our team for customised solutions.

No, there are no restrictions and you can upload as many files as you want.

Evidence levels in GoSpeech indicate how certain the AI model is that it has understood a word correctly. In GoSpeech, users can display the evidence levels according to their preferences.

Go to your list of recordings and open the desired file. Then click on the title of the recording - this will open a separate menu in which you can set the evidence level using the slide switch.

A word is highlighted in red if the AI model is unsure whether it has recognized the word correctly (this is only the case if the evidence levels are switched on). This marking helps users to quickly identify and check potential errors.

You must set the threshold value for the evidence level higher.

No, but the word may have different spellings or words with a similar phonetic stem.

If you have set the evidence level to 100%, words that have already been corrected manually are no longer marked, which is a good control mechanism.

If you activate the evidence level function, the words are only marked if the AI model has uncertainties when recognizing them. These markings are not permanent, but dynamic. This means that they can change when you adjust the evidence level settings. The markers are completely removed when the evidence levels are set to 0. The markers are not included in an export.

The context setting allows you to select a specific context (recording situation) when uploading a file, e.g., “conversation,” “dictation,” or “legal.” This optimizes speech recognition for the respective context and delivers more accurate transcription results. You can also control contexts via our API.

The context setting improves recognition accuracy by taking into account the linguistic characteristics of the selected context. For example, in dictations, punctuation marks and paragraphs can be spoken directly, while in the legal context, you get the correct legal spellings.

“Conversation” is the default context and is suitable for general dialogues, interviews, sessions, or meetings. This setting corresponds to the previous recognition in GoSpeech before the introduction of context settings.

Select “dictation” if you want to transcribe structured texts with spoken punctuation marks and paragraphs—e.g., for letters. GoSpeech also recognizes spoken commands such as “new paragraph” or “period.”

If you don't select a context, the context you last used will be reused. We recommend checking the selected context before each transcription.

Yes, choosing an inappropriate context can affect the quality of the transcription. Therefore, be sure to select the context according to the content of your recording.

Yes, we have a short explanatory video about the context setting feature. You can find it here: 👉 Watch the video on YouTube

“Legal conversation” refers to a conversation between several people in a legal context. This could be a conversation with a client or an audio recording from a courtroom, for example. “Legal dictation,” on the other hand, refers to a classic dictation in a legal context. 

Open your desired recording in the GoSpeech editor. Make sure that both the transcript and the notepad are visible. This allows you to start creating and linking notes straight away.

Select a relevant passage in the transcript and drag & drop it into Notepad. Alternatively, you can also use Copy & Paste. The note is automatically linked to the corresponding passage in the audio track.

Use the key combination ALT + left-click (under MacOS: option + left-click) on a word in Notepad. This will take you directly to the linked position in the audio track and in the transcript.

Yes, you can transfer individual words as well as smaller sections or entire speaker segments into Notepad - completely flexibly according to your needs.

Yes, you can edit or add to your notes at any time. However, please note: If you add or delete words manually, the automatic link to the audio track for these words will be lost.

Note synchronization is part of the GoSpeech Editor and is available to all users.

Yes, we have a short explanatory video on synchronizing notes with the audio track. You can find it here: 👉 GoSpeech: Synchronisation ihrer Notizen mit der Tonspur

Our transcription software excels at remarkably accurate recognition of audio and video files, even in complex scenarios with different speakers and different recording situations. Please note, however, that the recognition result depends on the quality of your recording. The better the source material is, the more accurate the result will be.

We are always at your disposal to help you with any difficulties and to find a suitable solution to your problem immediately. You can rely on our reliable support.

To conduct a live recording: 

1. Click the “RECORDING” button in the worklist.  

2. The “START NEW RECORDING” window will open.  

3. Fill in the required fields.  

• Mandatory fields: “Record Name” and “Title” are automatically filled with today’s date but can be changed.  
• Optional fields: Number/Reference and Producer.  

4. In the Recognition Settings, you can: 

• select the microphone for recording,  
• choose the delay for live transcription (more delay means better recognition; possible range is 0.7 to 4 seconds – at least 3 seconds is recommended),  
• select the recognition language,  
• choose the context: it is very important to select the correct context. (Contexts “Legal Conversation” and “Legal Dictation” are currently only available in German),  
• add vocabulary, which will be considered during the live recording,  
• choose whether one or multiple speakers are participating in the conversation. This selection will be saved for future recordings.  

Note: You may need to allow access to the selected microphone. 
5. Click “PROCEED.”  

6. A new recording will open. Set your preferred display for the live transcription. The display can be adjusted (size, color, language, microphone).  

7. Click “START” to begin recording.  

8. Speak as usual.  

9. Click “STOP” to end the recording.  

After stopping the recording, GoSpeech automatically generates a transcription. 

Yes, it is possible to pause a recording for up to 1 hour. If it is not resumed within 1 hour, it will be automatically stopped. 

To conduct a live recording: 

1. Click the “RECORDING” button in the worklist.  

2. The “START NEW RECORDING” window will open.  

3. Fill in the required fields.  

• Mandatory fields: “Record Name” and “Title” are automatically filled with today’s date but can be changed.  
• Optional fields: Number/Reference and Producer.  

4. In the Recognition Settings, you can: 

• select the microphone for recording,  
• choose the delay for live transcription (more delay means better recognition; possible range is 0.7 to 4 seconds – at least 3 seconds is recommended),  
• select the recognition language,  
• choose the context: it is very important to select the correct context. (Contexts “Legal Conversation” and “Legal Dictation” are currently only available in German),  
• add vocabulary, which will be considered during the live recording,  
• choose whether one or multiple speakers are participating in the conversation. This selection will be saved for future recordings.  

Note: You may need to allow access to the selected microphone. 

5. Click “PROCEED.”  

6. A new recording will open. Set your preferred display for the live transcription. The display can be adjusted (size, color, language, microphone).  

7. Activate “Virtual Mode.” 

8. Click “RECORD.” 

9. In the dialog that appears: 

• Select “Entire Screen” and, if you have multiple screens, choose the one where GoSpeech Medical is displayed. 
• Enable “Share system audio.” 
• Click “Share.” 

10. Click “START” to begin recording.  

11. Speak as usual.  

12. Click “STOP” to end the recording.  

After stopping the recording, GoSpeech automatically generates a transcription. 

Online live recordings can be recorded using Google Chrome and Microsoft Edge.

Yes, you can share live recordings with users in your organization so that they can: 

• view recordings  

• listen to the audio  

• export the transcript  

Users with editor permissions can also: 

• edit the transcript  

Note: 
To avoid conflicts, a recording can only be edited by one user at a time. 

Currently, only one live recording can be performed at a time. If multiple live recordings need to be processed in parallel, additional hardware resources are required. Each live recording additionally requires: 

• 1 CPU 

• 5 GB RAM 

• 3–4 GB of storage 

A configuration parameter can be used to control the number of parallel live recordings per installation and adapt it to the available resources. 

If these requirements are met and you need multiple recordings, please contact the service team.