Reviews from AWS Marketplace
0 AWS reviews
-
5 star0
-
4 star0
-
3 star0
-
2 star0
-
1 star0
External reviews
External reviews are not included in the AWS star rating for the product.
Best transcribing and audio APIs I've ever used
What do you like best about the product?
Deepgram knows who their customers are, developers or tech decision-makers in a company, so their site is made for them. It is so easy to understand everything, implement it quickly in any app and easy to find all information in the Documentation. I would use again and recommend it to others.
What do you dislike about the product?
Because it is made to be an abstract API as possible, to work for everyone, sometimes it takes some time to tweak all the options to work for a specific use case. In a way this is good because it gives control, but some pre-made recipe options would be helpful.
What problems is the product solving and how is that benefiting you?
I use Deepgram in 2 projects so far: a meeting-related product that transcribes stored audio/video recordings, and Deepgram is by far the best transcribing service. In another project, it is transcribing user intents in real-time, and is also really good at this too.
- Leave a Comment |
- Mark review as helpful
Best pricing, solid reliability and performance
What do you like best about the product?
- I couldn't find any other providers that did high quality transcriptions at that price.
- Never had any reliability problems.
- The service didn't skip a beat as I scaled.
- Large files transcribe very quick.
- Real time transcription works well.
- API was easy to use.
- Never had any reliability problems.
- The service didn't skip a beat as I scaled.
- Large files transcribe very quick.
- Real time transcription works well.
- API was easy to use.
What do you dislike about the product?
- Their propriatary Nova models aren't as good as whisper in my opinion (but they offer whisper as well so not a deal breaker).
- Related to that, there is not Whisper based real time api.
- Related to that, there is not Whisper based real time api.
What problems is the product solving and how is that benefiting you?
Deepgram allows me to transcribe audio at scale without having to worry about dev ops. This allows me to focus on my core product.
One of the best Speech to Text Ecosystems in the Market
What do you like best about the product?
Deepgram has been one of the best speech to text ecosystems for us. We have used almost all the models but Deepgrams stands out due to eas of integration, support for propeirotry as well as open source models (Whisper) and a great dashboard that allows us to track our usage transperantly.
What do you dislike about the product?
The NOVA model supports 19 languages as of now, we would love to see more languages get added to the model
What problems is the product solving and how is that benefiting you?
Deepgram is our voice/speech to text partner. We transcribe audio recordings and files using deepgram. We use deepgram's own models or Whisper via deepgram.
Powerful speech recognition than anything we've used
What do you like best about the product?
We're pleasantly surprised at the accuracy of Deepgram's speech recognition, and their supportive team that's responsive when we have questions.
What do you dislike about the product?
I wish there were more supported languages for their most advanced program Nova, but I know this is just a working progress and they are launching new languages on a rapid pace.
What problems is the product solving and how is that benefiting you?
As a language app focused on speaking, they are helping us deliver precise and accurate speech recognition and feedback to our users.
Loving the tranlation
What do you like best about the product?
We evaluated the major options for speech-to-text transcription and Deepgram was by far the best. Accuracy is really what we care about, and that's where Deepgram sjines.
What do you dislike about the product?
Deepgram does what we want it to do, no complaints.
What problems is the product solving and how is that benefiting you?
We recieve audio content that we want transcribed.
Very high quality transcriptions and excellent api
What do you like best about the product?
It was incredibly easy to access the deepgram API with Python. The documentation is pretty good and the structure of the data is fairly easy to work with. There is a lot there which is helpful for many different use cases. I was coding a solution that would take audio from hundreds of videos and send it to Deepgram to generate a transcription. It worked very very well overall.
What do you dislike about the product?
Honestly I can't think of anything. I love Deepgram. Very glad they are around.
What problems is the product solving and how is that benefiting you?
Automatic transcription of audio and video.
Hackathon Winner
What do you like best about the product?
I was involved in a Hackathon where the goal was to provide realtime translation in a setting like a church service to participants who were not fluent in the language being spoken. We realized pretty quickly that the most critical piece of accomplishing this was to have accurate transcripts from the original audio stream - without that the project was doomed. After a bit of research, we decided to use Deepgram due to its ease of integration, the configurability, and the ability to work with multiple input languages. There also were quite a few helpful examples and tutorials to get us started quickly. We ended up accomplishing our goals with Deepgram and ended up winning the Hackathon with our project.
What do you dislike about the product?
I was disappointed that there is currently no Java SDK or OpenAPI specification. It can be integrated with a Java application, but not via an SDK. We also noticed that in a church context, some words consistently were being incorrectly transcribed. For example, Ephesus was being transcribed as emphasis even though it didn't fit the context of the sentence.
What problems is the product solving and how is that benefiting you?
Producing high quality transcripts from an audio stream so that we can provide realtime translations.
A smooth experience but they can get even better
What do you like best about the product?
They are quite simple to integrate in the project, and have a great dashboard where a lot of the info is accessible.
There is a generous trial plan for quite a lot of testing and enough to kickstart you if you're a small project.
There is a generous trial plan for quite a lot of testing and enough to kickstart you if you're a small project.
What do you dislike about the product?
I want to be able to download older transcripts / results.
The support can get a bit better with less ambiguous answers.
The support can get a bit better with less ambiguous answers.
What problems is the product solving and how is that benefiting you?
They have an api integration for whisper transcription, which we are using for transcript and diarization of audio recordings.
Fast and realtime voice transcription
What do you like best about the product?
Speed and realtime results at competitive prices. Very easy to setup and good overall documentations.
What do you dislike about the product?
The detection is still not as good as OpenAI whisper with the post processing that seems to make it better than Deepgram.
The number of different models and config combinations can be a bit intimidating to arrive at the best set for the particular use case.
The number of different models and config combinations can be a bit intimidating to arrive at the best set for the particular use case.
What problems is the product solving and how is that benefiting you?
We need realtime voice recognition to have a conversational AI app
Fastest STT for our use-case
What do you like best about the product?
Deepgram's speed in converting spoken words to text is truly impressive. This real-time capability allows us to process and transcribe conversations quickly, making it a valuable tool for our operations.
User-friendly interface and API integration make it easy for developers.
Diarization feature is a game-changer. It helps us differentiate between multiple speakers in a conversation, making our transcriptions more organized and meaningful.
Support for telephony codecs, such as Mu-law 8000, is essential for us, as we deal with a variety of audio sources. This versatility ensures we can work with different types of recordings seamlessly.
The availability of a phone call model is a great asset for our use case. It's tailored to handle telephone conversations, resulting in more accurate transcriptions for this specific scenario.
The ability to receive interim results is a big plus. It allows us to get an early glimpse of the transcription, which can be useful for real-time applications. Additionally, the endpointing feature helps us identify the start and end of speech segments accurately.
User-friendly interface and API integration make it easy for developers.
Diarization feature is a game-changer. It helps us differentiate between multiple speakers in a conversation, making our transcriptions more organized and meaningful.
Support for telephony codecs, such as Mu-law 8000, is essential for us, as we deal with a variety of audio sources. This versatility ensures we can work with different types of recordings seamlessly.
The availability of a phone call model is a great asset for our use case. It's tailored to handle telephone conversations, resulting in more accurate transcriptions for this specific scenario.
The ability to receive interim results is a big plus. It allows us to get an early glimpse of the transcription, which can be useful for real-time applications. Additionally, the endpointing feature helps us identify the start and end of speech segments accurately.
What do you dislike about the product?
While Deepgram's accuracy is commendable, there is room for improvement in the Word Error Rate. In real-time conversations, every word counts, and even a slight inaccuracy can be problematic. Although it's acceptable for our use case, further improvements in this aspect would be highly beneficial.
What problems is the product solving and how is that benefiting you?
Deepgram is benefiting our organization by providing an efficient, cost-effective, and feature-rich hosted STT solution. It eliminates the need for us to make substantial investments in building and maintaining an in-house STT system, allowing us to focus on our core business activities and providing us with reliable, accurate, and scalable transcription services. This has not only saved us time and resources but also improved the overall quality of our real-time conversation-based applications.
showing 11 - 20