Amazon Transcribe Medical

Digitization of the healthcare sector

In recent years, the healthcare sector has begun to actively embrace modern digital solutions - from telemedicine applications, connecting residents of the most remote, “hard-to-reach” regions to world-class medical services, to the use of sensors and devices that help remotely monitor and record patient physical data such as: heartbeat, blood pressure, movement and even behavioral patterns. The unique challenges of Covid-19 have played a decisive role in accelerating the digitization of healthcare, when it became clear that many processes in the healthcare sector require a fundamental transformation.

Currently, medicine has a variety of digital tools to improve communication, administrative and operational processes, data storage and transition.

One such tool that facilitates the work of medical professionals is Transcribe Medical Service from Amazon.


What is Amazon Transcribe Medical?

In the past, writing paper reports took doctors a lot of time. And after the beginning of digital transition there is a standard requirement for healthcare providers to enter medical records into Electronic Health Record (EHR) systems on a daily basis. According to a study held by the University of Wisconsin and the American Medical Association in 2017, primary care physicians in the United States spent up to 6 hours a day entering this data.

In 2019, Amazon launched a service built on top of the Amazon Transcribe. It was designed specially for healthcare professionals to transcribe medical-related speech, such as physician-dictated notes, drug safety monitoring, telemedicine appointments and consultations, or conversations of doctors with patients. 

The Amazon Transcribe Medical service uses machine learning and natural language processing (NLP) to accurately convert audio speech or conversation to a text. It is trained to understand complex medical language and special terms and measurements used by doctors. Developers can use Amazon Transcribe Medical for medical voice applications, by integrating with the service’s easy-to-use APIs. Pharmaceutical companies and healthcare providers can use Amazon Transcribe Medical to create services that enable fast and accurate documentation. 

The service can transcribe speech as either an audio file or a real-time stream, the input audio can be in FLAC, MP3, MP4, Ogg, WebM, AMR, or WAV file format. Streaming transcription is available in US English, it can produce transcriptions of accented speech, spoken by non-native speakers.

This service provides transcription expertise for primary care and specialty areas such as cardiology, neurology, obstetrics-gynecology, pediatrics, oncology, radiology and urology. Transcription accuracy can be improved by using medical custom vocabularies.


Transcribe Medical use cases

Medical dictation: medical specialists can record their notes by speaking into the microphone of a mobile device during or after interacting with a patient, being able to reduce the administrative workload and focus on providing quality patient care.

Drug safety monitoring: transcribing of phone calls regarding drug appointment and side effects enables more safe service provisioning by pharmaceutical companies and clinics. 

Transcribing of conversations: recording conversations between a doctor and a patient in real time without disrupting the interaction, allows healthcare providers to accurately capture details such as mentioned symptoms, medicine dosage and frequency, side effects. This information can be processed through subsequent text analytics and then entered into Electronic Health Record (EHR) systems.

In case of online video or phone consultations Channel Identification feature can be used. This is a powerful tool allowing to independently transcribe the patient and clinician audio channels and provide real-time conversational subtitles.


Benefits of Amazon Transcribe Service

Amazon Transcribe Medical benefits a wide range of healthcare specialists: nurses, physicians, researchers, insurers, and pharmaceutical companies - as well as their patients. The following features make it highly attractive to clinicians and healthcare professionals:

HIPAA (Health Insurance Portability and Accountability Act) eligible: providing support for the automatic identification of protected health information (PHI) in medical transcriptions Amazon Transcribe Medical reduces the cost, time, and effort expended on identifying PHI content through manual processes. PHI entities are labeled clearly with each output transcript, making it convenient to build additional downstream processing for a variety of purposes, such as redaction prior to text analytics.

Highly accurate transcription: the narrow specialization of the service, exclusively aimed  at the needs of the healthcare sector, ensures that even the most complex medical terms, such as the technical names of diseases and medicines, are recorded correctly. 

Improving the patient and practitioner experience: so that the doctor does not have to waste time taking notes and writing reports, but can focus on the patient, accurately transcribing all the details of the consultation or conversation without disrupting the interaction.

Easy to use: no prior knowledge or experience with machine learning is required. Developers can focus on building their medical speech applications by simply integrating with the service's APIs. Transcribe Medical handles the development of state-of-the-art speech recognition models.

Thus, Amazon keeps investing into the medical sector, empowering healthcare and life sciences, and enhancing the number of digital services to deliver patient-centered care, accelerate the pace of innovation and unlock the potential of data, while maintaining the security and privacy of health information.


Our experience

With extensive experience in building healthcare applications based on Amazon services and developing long-term partnership with global leaders in telemedicine technologies & services, we, Inmost Company, took the opportunity to ease the burden of reporting and documentation for our clients by integrating Transcribe Medical into the application for remote medical consultations. This has significantly optimized medical staff workload, streamlined processes, and increased positive feedback from patients.

Based on these experiences, we consider Amazon Transcribe Medical Service to be a really important and useful tool for transforming medical services. 

And, of course, we are ready to support healthcare organizations on their digital transformation path by providing consulting services, renovating and improving existing platforms or developing efficient and reliable solutions from scratch.



According to Statista the quantity of IoT devices is about 43 billion now.

And by 2025, 75 billion IoT devices are predicted to be online and Statista predicts that rather a lot of those devices will be in areas that lack a standard connection.

The future of IoT will be built through open networks and collaboration. Until the future has not come, let's discuss the variants of connection for nowadays.

I think there is no need to mention BLE, Wi-Fi, or 5G. There is no competition between these networks – rather, they are complementary.

Let’s speak about Zigbee. What is this technology different from above mentioned?

Zigbee and "what it is eaten with"

Zigbee is a standards-based wireless technology developed as an open global market connectivity standard to address the unique needs of low-cost, low-power wireless IoT data networks. The Zigbee connectivity standard operates on the IEEE 802.15.4 physical board radio specification and operates in unlicensed radio bands including 2.4 GHz, 900 MHz and 868 MHz.

Specifications of Zigbee

The Zigbee specifications, which are maintained and updated by the Zigbee Alliance, boost the IEEE 802.15.4 standard by adding network and security layers in addition to an application framework.
In theory, it enables the mixing of implementations from different manufacturers, but in practice, Zigbee products have been extended and customized by vendors and, thus, plagued by interoperability issues. In contrast to Wi-Fi networks used to connect endpoints to high-speed networks, Zigbee supports much lower data rates and uses a mesh networking protocol to avoid hub devices and create a self-healing architecture.

There are three Zigbee specifications: Zigbee PRO, Zigbee RF4CE and Zigbee IP.

Zigbee PRO aims to provide the foundation for IoT with features to support low-cost, highly reliable networks for device-to-device communication. Zigbee PRO also offers Green Power, a new feature that supports energy harvesting or self-powered devices that don't require batteries or AC power supply.

Zigbee RF4CE is designed for simple, two-way device-to-device control applications that don't need the full-featured mesh networking functionalities offered by the Zigbee specification.

Zigbee IP optimizes the standard for IPv6-based full wireless mesh networks, offering internet connections to control low-power, low-cost devices.

Mesh network

Mesh networks are decentralized in nature. It’s flexible, reliable and expandable - End Node, Router or Coordinator, where nodes can communicate peer-to-peer for high speed direct communication, or node to Gateway.

Zigbee and Z-wave are two well-known mesh networking technologies. In a mesh network, nodes are interconnected with other nodes so that multiple pathways connect each node. Connections between nodes are dynamically updated and optimized through sophisticated, built-in mesh routing tables.


Zigbee is inherently secure. It provides options for authentication and data encryption. Zigbee uses 128-bit AES encryption keys, similarly to its primary competitor, Z-Wave (all pros and cons of Z-wave will be considered in the next article).
This plus short-range signals make Zigbee secure. However, most home automation protocols have similar levels of security when you configure them properly. 

Power consumption

Power consumption for Zigbee is comparable with BLE. However, the proven, routed mesh mechanism adopted in Zigbee makes it slightly more power efficient.

What Is Zigbee Compatible With?

The devices are controlled by Samsung SmartThings and Zigbee. Amazon Echo Dot,  Philips Hue, IKEA Tradfri. Hive Active Heating is a device that uses natural gas and has accessories. Honeywell manufactures a variety of thermostat products.

Conclusion. Why choose Zigbee?

Comparing Zigbee with existing variants of connections, it’s obvious that Zigbee offers multiple advantages over Bluetooth.

For example, BLE works best for smaller size packets (i.e. less than 12 bytes). For smaller size (less than 12 bytes), its comparable to Zigbee but as packets size starts increasing BLE higher layers do the fragmentation and cause latency to increase.

However, BLE has a cost advantage over Zigbee. BLE mesh has a bigger eco system and uses the same BLE chipset used in other applications, therefore high scale production of BLE chipsets pulls down the cost of IC compared to Zigbee.

Need of gateway device for Zigbee further increases the cost of the overall system. BLE based systems can provide limited functionality (everything except full-fledged internet connectivity) without a gateway as well. In addition, licensing of Zigbee is more expensive and complex than BLE.

Meanwhile, Zigbee is more cost-effective and uses significantly less energy than Wi-Fi, resulting in better battery life. To speak about another “rival” LoRaWAN, it’s significantly cheaper than Zigbee and  they are close by some characteristics.

And if you are looking for a cheap and  long battery life sensing project, where no real-time, control or automation requirements are anticipated and slower poll-rates are suitable, then LoRaWAN is a good contender and is a good choice for many entry-level sensing applications.
But, if it is necessary to control automation or faster poll rates, it’s better to step up to Zigbee. As it was mentioned Z-Wave will be considered next time.

Where to use?

The Zigbee wireless communication system is used by homes, businesses, and other locations to communicate.
Zigbee can transmit data over a long distance, which is sufficient for most applications. Zigbee is a clear winner for industrial applications that require reliability, real-time monitoring, control or automation and this protocol is highly under-rated for low power sensing.



NFT – the most contradictory component of the Web 3

There are probably no people left who have not heard of NFTs yet.

Being a vital component of Web-3: the next iteration of the Internet, along with the Metaverse and De-Fi, NFTs evoke, perhaps, the most contradictory feelings in society - from enthusiasm, sometimes bordering on insanity, to outright hostility and harsh criticism.

NFT - a non-fungible token - is a unique unit of data that is verified and stored in the blockchain and can be linked to digital or physical assets to provide immutable proof of ownership. Blockchain technology allows NFTs to be tracked in an immutable digital ledger that provides a history of assets and can be verified at any time. So, NFTs cannot be replicated, destroyed, or counterfeited.

NFTs are primarily created on Ethereum, but other blockchains support them as well.

For selling NFTs, they must first be minted. Minting an NFT means converting a digital file into a digital asset that can be published and stored on the blockchain, making it available to potential buyers. The minting process is not free - you need a crypto wallet and a certain amount of crypto currency to cover the Ethereum "gas fees." The most popular NFT marketplaces on the Ethereum blockchain are: OpenSea NFT, Rarible, and Mintable.

Today, almost anything can become an NFT: paintings, photos, videos, music, gifs, memes - any kind of unique art that can be represented digitally. Or it can even be real estate, collectibles, event tickets, website domains or tweets.

Famous auction houses Christie’s and Sotheby’s have already made sales of NFT artworks for several hundred million dollars.

The first experiments with NFT started back in 2013, but the wave of hype rose only in 2021, and sometimes it looked like real insanity. So, the creator of the Nyan Cat meme received $580,000 in cryptocurrency for a gif with the famous cat meme, and the digital artist Beeple sold the token of jpeg collage - Everydays: The First 5000 Days, for $69.3 million.

NFT technology is actively used by both well-known and not yet recognized artists. The main factor in the growing popularity of NFT is the opportunity for a beginner to present his or her work to a wide audience. A few years ago, a new artist had to work hard for several years before reaching the first serious exhibition, and still success was not guaranteed. Today it's enough to convert your painting into a digital format, create a corresponding NFT token (it's not that complicated) and sell it for real money.

Actually, NFT technology can be used for transactions with any digital assets, however, the recent trends show a growing interest in selling real things as NFTs. These can be, for example, sculptures, antiques, a coin collection, etc. But if converting paintings into a digital format is a common thing today, how can a real physical object become an NFT?

The first way is to create a 3D digital copy of the object. Technologies that allow the average person to create such copies are becoming more and more available. And of course, it attracts a lot of interest from businesses and corporations that have been already using and investing in 3D technologies to promote their brands and products not only in the real world, but also in the virtual world of the metaverse. In addition, NFTs of 3D objects are expected to replace our favorite real-world things, objects, and assets in the metaverse, making it even more similar to our everyday environment.

But what if we take objects that are difficult to digitize? Last month, for example, a three-bedroom house in South Carolina was sold as an NFT for $175,000. The buyer indicated that he was able to make the transaction for that property with just one click. How does it work?

In simple terms, the selling company creates an NFT that represents ownership of the house. Those who buy this NFT become the owner of the property. Despite the fact that the purchase is made digitally, the ownership is considered absolutely real - whoever owns the NFT owns the house in the real world.

Although such transactions are still viewed with suspicion by the majority, there are serious reasons to believe that NFT technology could open the door to a decentralized economy without intermediaries such as banks or a government. In the future, it may completely change the rules of the markets.

Today, besides the arts and real estate, the most potential for NFTs have gaming, education,  healthcare, supply chain and logistics industries. NFT tokens can be used to confirm any important document: a diploma, health records, a marriage registration certificate, etc.

A serious barrier to NFT adoption into the mentioned areas is the lack of government regulation. And it is very likely that in case of a fraud or hacker attack, the affected party will not be able to recover its losses.

Moreover, NFT technology faces many other challenges today. For example, one of the main arguments against NFT is its huge energy consumption and extremely negative impact on the environment. However, after Ethereum has switched from a power-intensive Proof-of-Work protocol to a mining-free Proof-of-Stake, it is possible that NFT will become more eco-friendly and increase its audience.

Since NFT has both - the devoted fans and haters, it remains one of the most hype-boosting components of web-3, and is mentioned every now and then in the news and social media, either along with the figures containing impressive number of zeros, or along with facts that cause a no less impressive number of questions and misunderstandings.



Transcribe Service was launched by Amazon in 2017 enabling developers to implement a speech-to-text feature to their applications.

Analyzing and data extraction from audio files is almost impossible for computers. To use such data in an application, speech must first be converted to text. Services performing speech recognition technologies have certainly existed before, but they were generally expensive and poorly adapted to various scenarios, such as low-quality phone audio in some contact centers.

Powered by deep learning technologies, Amazon Transcribe is a fully managed and continuously trained automatic speech recognition service that automatically generates time-stamped text transcripts from audio files. The service parses audio and video files stored in many common formats (WAV, MP3, MP4, AMR, Flac, etc.) and returns a detailed and accurate transcription with timestamps for each word, as well as appropriate capitalized words and punctuation. For most languages, numbers are transcribed into a word form, however for English and German languages Transcribe treats numbers differently depending on the context in which they're used.

Now Transcribe supports 37 languages.

Transcription methods can be divided into two main categories:

  • Batch transcription: transcribing media files that have been uploaded into an Amazon S3 bucket;
  • Streaming transcriptions: Transcribe media streams in real time.


Here are some of the features it provides:

  • Single and multi language identification: identifying the dominant language spoken in your media file and creating a transcript. If speakers change language during a conversation, or if each participant speaks a different language, your transcription output correctly detects and transcribes each language;
  • Transcribing multi-channel audio: combines transcriptions from multi channel audio into a single output file. It is possible to enable channel identification for both batch processing and real-time streaming;
  • Speaker diarization: the partition of the text from different speakers, detecting each speaker in the provided audio file;
  • Custom language models: designed to improve transcription accuracy for domain-specific speech. This includes any content that goes beyond the everyday type of conversations. For example, an audio recording of a report from a scientific conference will obviously contain special scientific terms that standard transcription is unlikely to be able to recognize. In this case, you can train a custom language model to recognize the specialized terms used in your discipline;
  • Custom vocabularies: are used to improve transcription accuracy for a list of specific words. These are generally domain-specific terms, such as brand names and acronyms, proper nouns, and words that Amazon Transcribe isn't rendering correctly;
  • Tagging: adding custom metadata to a resource in order to make it easier to identify, organize, and find in a search;
  • Subtitles: can be used to create closed captions for your video and filter inappropriate content from your subtitles.


Transcribe offers indispensable features for call centers and support services. It helps to capture useful insights by transcribing customer calls in real time. Analyzing and categorizing calls by keywords, phrases and sentiment can help track negative situations, identify trends in customer issues or allocate calls to specific departments.

It is possible to measure the volume of speech. This metric helps to understand if the customer or employee is talking loudly, which is often an indication of being angry or upset. The quality of communication with the client can also be determined by setting the following metrics: interruptions, non-talk time, talk speed, talk time.

Besides call-centers, Transcribe Service can be useful in almost any field: education, law, e-commerce, and many others. For example, Amazon Comprehend Medical is a machine-learning-powered HIPAA-eligible service pre-trained to identify and extract health data from medical texts, such as prescriptions, procedures, or diagnoses. 

It is difficult to imagine modern technologies without a service that can transform speech into text. And of course, Transcribe has analogues from other digital giants. However, it is worth noting that a large number of developers who have leveraged Amazon service, admit a much higher quality and accuracy compared to similar solutions provided by the current market.