Ultimate Guide to Speech to Text APIs in 2023

Ultimate Guide to Speech to Text APIs in 2023

This article includes a comprehensive guide to assist you in choosing the best speech-to-text provider for your needs among the numerous options available.

Navigating the various speech-to-text providers and understanding their unique offerings can be challenging, but this guide aims to simplify the selection process and provide you with the information you need to make an informed decision, saving you time and effort:

Eden AI - Ultimate Guide to Speech-to-Text APIs

What is Speech-to-Text?

Speech-to-Text (STT) technology allows you to turn any audio content into written text. Also known as Automatic Speech Recognition (ASR) or computer speech recognition, Speech-to-Text is based on acoustic modeling and language modeling. 

Speech-to-Text APIs use cases

You can use Speech Recognition in numerous fields, and some STT APIs are built especially for those fields. Here are some common use cases:

  • Call centers: data collected and recorded by speech recognition software can be studied and analyzed to identify trends in customer
  • Banking: to make communications with customers more secure and efficient.
  • Automation: to fully automate tasks like appointment bookings or finding out where your order is
  • Governance and security: to complete an identification and verification (I&V) process with the customer speaking their details such as account number, date of birth, and address.
  • Medical: for voice-driven medical report generation or voice-driven form filling for medical procedures, patient identity verification, etc
  • Media: automated process for TV, radio, social networks videos, and other speech-based content conversions into fully searchable text.

A well-supplied Speech-to-Text market

There are many companies in the speech recognition market, both large and small, that offer various strengths and weaknesses. 

Some of the major players in the field include Google Cloud, Amazon Web Services (AWS), Microsoft Azure, and IBM Watson, which offer highly accurate and performant generic speech-to-text APIs. These companies have trained their models on large amounts of data to achieve their high levels of accuracy. 

There are also Speech-to-text specialized companies that provide highly effective Speech-to-text APIs: Rev AI, Assembly AI, Deepgram, Speechmatics, Vocitec, Symbl.ai, NeuralSpace, Amberscript, Speechly, etc. All those providers can be particularly efficient for specific languages, offer specific features, or support specific file formats.

Speech-to-Text providers available

Eden AI’s Ultimate Guide to choose the best STT API

It can be challenging to navigate the many speech-to-text providers and understand their unique offerings. That's why Eden AI's speech experts have created an ultimate guide to help you make an informed decision and save time when selecting a supplier. The guide is divided into four aspects:

  • Features: All the different options provided by the speech-to-text APIs. These options can improve the quality of the transcription and also provide you with more information in the response. These features are outlined in further detail below.
  • Language supported: It's important to consider the languages supported by the provider, as some may specialize in specific accents or have the ability to handle rare and exotic languages. 
  • File format supported: Audio files can be encoded in a variety of formats, beyond the commonly known .mp3 and .wav formats. Some of these encoding use either a lossy or lossless compression, supporting each a range of sample rating with a bit-depth precision usually around 16-bits /24-bits.
  • Pricing: Speech-to-text APIs can have different prices from simple to double, based on volume levels, license, or per request.

This guide was created by Eden AI's speech-to-text experts in collaboration with participating providers. It includes all of the necessary information for choosing a speech-to-text supplier. Eden AI maintains a neutral stance and does not have any interest in promoting one supplier over another.

What features should you consider for your Speech-to-text transcription?

Speech-to-text technology provides a great deal of additional information and analysis beyond simply transcribing the audio. In many cases, users need more detailed information to extract valuable insights from the audio content. Here are some examples of the types of information that can be included in a text-to-speech API response:

Standard features

Speaker Diarization

Speaker diarization is the process of segmenting audio recordings by speaker labels and aims to answer the question “who spoke when?”. In the Automatic Speech Recognition field, Speaker diarization refers specifically to the technical process of applying speaker labels (“Speaker 1”, “Speaker 2”, etc.) to each utterance in the transcription text of an audio/video file.

Here is an example of a transcription without speaker diarization on Eden AI platform:

Here is the same example with speaker diarization:

Speaker diarization involves multiple tasks:

  • Detection: divide speech and noise
  • Segmentation: split audio file into small segments
  • Embeddings representation: all those segments are converted into vectors by a Neural networks
  • Clustering of those embedding: each cluster corresponds to a speaker


Most of the speech-to-text APIs return timestamps in their response. The timestamps may be provided "per word" or "per phrase", depending on the API. These timestamps can be useful for synchronizing transcriptions with the audio, or for identifying specific points in the audio for further analysis.

Automatic language detection

No need to set the language of the audio file in your request, it can be automatically detected by some STT APIs. This can save time and money, as it eliminates the need to use a separate language detection API before the speech-to-text process. 

Using an STT API with integrated automatic language detection can also reduce latency, compared with two APIs calls  (one for language detection API, then one for speech-to-text).


Some speech-to-text APIs automatically add punctuation to the transcription. This feature can be particularly useful for generating subtitles, as it helps to make the transcription more readable and understandable. The addition of punctuation can also improve the usability of the transcription by providing a clearer structure and better organization of the spoken content. 

Profanity filter

Speech-to-Text can automatically detect profane words in your audio data and censor them in the transcript. This avoids you to use text explicit content detection behind your Speech-to-text API request.

Noise filter

Many speech-to-text APIs include a noise filter to help improve transcription accuracy in real-world environments where the audio may be contaminated with background noise. In these situations, the API must be able to distinguish between spoken words and noise, and a noise filter can help to reduce the impact of noise on transcription accuracy. 

This is especially important when the audio quality is poor, as transcription accuracy can suffer without the help of a noise filter. By reducing the impact of noise on transcription, a noise filter can help to improve the overall accuracy and usefulness of the transcription.

NLP analysis: keyword, NER, sentiment, summarize, etc

Some Speech-to-text APIs can extract additional information from the transcript such as: keywords, entities, sentiment, emotions, etc.. You can also have a translation or summarization of the transcript. Those options can sometimes imply an extra cost. If you don’t have good performance with integrated NLP analysis, you can still use NLP APIs from Eden AI after your Speech-to-text API request.

Custom features

Some speech-to-text API providers allow users to include optional parameters in their requests in order to help improve the accuracy of the transcription. 

Specific domain

Some speech-to-text API providers offer the option to select a specific enhanced model that has been specifically trained for a particular type of audio, such as medical conversations, financial discussions, meetings, or phone calls. 

By using a model that has been specifically designed for a particular field, users may be able to achieve higher levels of accuracy and more relevant transcriptions.

Custom vocabulary 

Some speech-to-text APIs provide a parameter that allows users to specify a custom dictionary of words to help improve transcription accuracy. This can be particularly useful for domain-specific terms, such as brand names, acronyms, and proper nouns, which may not be recognized by the API's general speech recognition engine.

Here is an example of custom vocabulary parameters on Eden AI platform:

Speech-to-text is available for every language.

Transcribe audio from all over the world:

Many speech-to-text APIs support transcription of audio in a wide range of languages, with some providers offering support for up to 250 different languages. Some providers may have a particular focus on certain regions or language groups, such as Asian languages, African languages, or European languages, while others may offer more comprehensive coverage. 

Additionally, some APIs may be able to transcribe audio in dialects or other variations of a given language.

Refine the transcription by choosing a model optimized for the country's accent

Some speech-to-text APIs offer the option to select a specific language region or accent when requesting transcription of an audio file. For example, a user may be able to choose between different variations of Spanish, Arabic, or English, depending on the API. You can choose between 24 different Spanish languages, 22 different Arabic languages, and 17 English languages.

Multiple formats available to parse in Speech-to-text APIs

Most of the Speech-to-text APIs support standard audio file formats such as .mp3, .wav, and .mp4 (video). Some providers also support other formats withossless compression: .flaac, .aac, etc…

For more specific use cases you might need to process your audio file with specific formats:

  • .speex specifically tuned for the reproduction of human speech
  • .aiff and .m4p are Apple audio file formats
  • .wma is Windows Media Audio developed by Microsoft

Benefits of using Speech-to-Text API with Eden AI

Using Speech-to-Text with Eden AI API is quick and easy.

Multiple AIs in one API - Eden AI

Save time and cost

We offer a unified API for all providers: simple and standard to use, with a quick switch that allows you to have access to all the specific features very easily (diarization, timestamps, noise filter, etc.).

Easy to integrate

The JSON output format is the same for all suppliers thanks to Eden AI's standardization work. The response elements are also standardized thanks to Eden AI's powerful matching algorithms. This means for example that diarization would be in the same format for every speech-to-text API call.


With Eden AI you have the possibility to integrate a third party platform: we can quickly develop connectors. To go further and customize your speech-to-text request with specific parameters, check out our documentation.

Want to use Eden AI?

Eden AI has been made for multiple speech-to-text APIs use. Eden AI is the future of speech recognition usage in companies. The Eden AI API speech-to-text APIs allows you to call multiple speech-to-text APIs and handle all your voice issues.

You can use Eden AI speech-to-text to access all the best STT APIs of the market with the same API. Here are tutorials for Python(lien) and JavaScript(lien).

‍The Eden AI team can help you with your speech recognition integration project. This can be done by :

  • Organizing a product demo and a discussion to better understand your needs. You can book a time slot on this link: Contact
  • By testing the public version of Eden AI for free: however, not all providers are available on this version. Some are only available on the Enterprise version.
  • By benefiting from the support and advice of a team of experts to find the optimal combination of providers according to the specifics of your needs
  • Having the possibility to integrate on a third party platform: we can quickly develop connectors

Related Posts

Try Eden AI for free.

You can directly start building now. If you have any questions, feel free to schedule a call with us!

Get startedContact sales