What do users say?

I like that the API is great, which has helped in building AI solutions and voice agents. The pricing is reasonable, and I appreciate the credits offered to get startups going. Deepgram has always been on the leading edge of transcription, which is crucial for me. I love the Flux tool; it's worked really well and has been game-changing in reliably detecting when people are done speaking. I like how stable Deepgram is, and it offers high-quality transcription at a reasonable price. Review collected by and hosted on G2.com.

I honestly have no complaints. I find the Deepgram API to be the easiest of the major vendors and it has been super stable. Review collected by and hosted on G2.com.

There are two things I particularly like about Deepgram. First, its determinism ensures that transcribing the same voice over on different API calls remains consistent with 99.9% accuracy, even down to the millisecond. I have translated at least 300 to 400 voice overs and have never seen any inconsistency. Every time I send a voice over, I receive the output with the same time stamp. The second thing is that its segment level time stamp feature remains grammatically correct. There are no issues or errors in punctuation and capitalization, and it also does segmentation very well. Review collected by and hosted on G2.com.

I didn't find anything in Deepgram that needs improvement, but one feature I would like to see in Deepgram is the ability to set granularity. If we are given the opportunity to set this in API calls or in some other way, so that we can specify how many words we want in a segment. Because sometimes if a sentence is very long, it combines segments of fifteen, twenty, or twenty-five words, which is very, very low. Review collected by and hosted on G2.com.

I used Deepgram for my voice agent project and really appreciate its low latency, streaming reliable transcripts for text-to-speech. It's helpful for real-time voice calls, making the experience feel smooth and natural. I also find it very easy to set up, especially using APIs. Compared to other platforms like Eleven Labs, Deepgram is not less costly but also provides better performance. Deepgram is invaluable when smoothness in real-time calls is a top priority, handling live calls more effectively by instantly converting conversations into text. Review collected by and hosted on G2.com.

Sometimes the multi-language support doesn't work perfectly. When calling customers who speak Indonesian or Dubai dialects, it doesn't detect their language well. Review collected by and hosted on G2.com.

I use Deepgram for building voice-enabled applications and handling speech-to-text in real time. I like that Deepgram has high accuracy for real-time speech-to-text. Smart formatting is great as it makes the output ready to use without extra processing, and keyword boosting is valuable for domain-specific use cases. Deepgram converts audio into text quickly and enables voice features without a complex speech recognition system. The initial setup is easy, straightforward, and based on a simple API key. Review collected by and hosted on G2.com.

I think no need much more improvement, it's strong but improve cost predictability and deeper customization. Cost - it helps to have usage estimation and alert to track spending in real time. Customization - easy ways to fine-tune models for specific domains. Review collected by and hosted on G2.com.

I use Deepgram for accurate and fast speech-to-text transcription in my application, which saves time by quickly converting audio into accurate text, reducing manual transcription effort. I like its accuracy and fast processing and how easy the API integration is. Its real-time transcription and reliable performance make it very efficient to use. The initial setup was very easy and quick to implement, and using it with tools like Python makes my workflow seamless. Deepgram's speed and cost efficiency are much better, which is why we switched from other tools. Review collected by and hosted on G2.com.

It could improve in handling heavy accents and reducing errors in noisy environments. Improving accent training and noise cancellation models would enhance accuracy in challenging audio conditions. Review collected by and hosted on G2.com.

I have used Deepgram for my Voice Agent software's Speech-To-Text (STT) and I like that it helps solve critical challenges like real-time speech-to-text processing even in noisy areas, and reduces latency for our application. What I really like about Deepgram is that it's fast, accurate, and offers low latency and real-time transcription, which makes it easy for developers to implement. The low latency is particularly beneficial for processing speech in real time, making AI conversations feel natural. The speed and low latency ensure smooth conversations between humans and agents, catching voices quickly and processing them efficiently. The initial setup was straightforward and quick, with very clear documentation. Review collected by and hosted on G2.com.

Overall working with Deepgram has been great. It should have little improvised where multiple voices or most noisy area, where it will catch the voice the nearest one for better streaming. Review collected by and hosted on G2.com.

I use Deepgram for my voice-agent project and really appreciate its low latency streaming and reliable transcripts for speech-to-text (STT). This is especially helpful for real-time voice calls, making the experience feel smooth and natural. I also find it very easy to set up, which is a big plus. Deepgram is not only less costly but also provides better performance with less latency. These aspects of Deepgram are invaluable when smoothness in real-time calls is the top priority. Review collected by and hosted on G2.com.

I think all things are very excellent but in the multi language support it is not supporting all the languages. I recently tried multi language option `language="multi"` in Deepgram Speech to text service and called an Indonesian customer but when the customer tried to talk in Indonesian language then it is not detecting. So I think it should be improved, rest all the features are very amazing. Review collected by and hosted on G2.com.

I like Deepgram for its speed, accuracy, and developer-friendly APIs when working with voice data. I appreciate the very low latency and high transcription accuracy, which improve the real-time experience and reduce manual work. The easy API integration is also great for real-time and batch processing. Furthermore, the developer-friendly APIs enhance productivity by making voice feature implementation quick, easy to maintain, scalable, and testable. The initial setup of Deepgram is straightforward, aided by quick account setup, simple SDK installation, easy transcription requests, and helpful documentation. Overall, the cost efficiency and better performance at scale compared to previous tools my team used make Deepgram very appealing. Review collected by and hosted on G2.com.

While Deepgram is a strong platform overall, there are a few areas where it could be improved to better meet expectations like accuracy with heavy accents, noisy environments, or specialized vocabulary, language support and multilingual conversations, and pricing predictability at scale. Additionally, actual speaker identification in multi-speaker conversations could be better. Review collected by and hosted on G2.com.

I use Deepgram as more than just a transcription tool; it's my gateway to turning raw audio into intelligence. I like how fast and real-time it is, which makes it really useful for live conversation, not just recordings. I also appreciate that it works well with my development setup and is easy to connect. The initial setup was fairly simple, and overall, it was easy to start using. We switched to Deepgram because it was faster and more accurate than the standard speech-to-text tool we used before. Review collected by and hosted on G2.com.

One thing that could be improved is accuracy with different accents or noisy audio. Also, the setup and fine-tuning can take a bit of time if you are new to it. Better handling of strong accents and background noise would make results more consistent across different real-world situations. Simplify the setup process. Review collected by and hosted on G2.com.

I like Deepgram's fast and highly accurate speech-to-text transcription. It processes audio in real time and handles different accents well, making the transcripts reliable. Real-time processing is useful for live conversations and meetings, as it provides instant transcriptions without waiting. Deepgram also makes it easy to build voice-enabled applications that can understand users clearly and respond quickly. The initial setup was very easy, with clear documentation, a simple API, and helpful examples that made integration smooth and quick. Using Deepgram allowed us to add voice transcription features without needing a large development team. Review collected by and hosted on G2.com.

Deepgram works very well overall, but I think it could improve in a few areas. For example, expanding support for more languages and improving accuracy in very noisy audio environments would make it even better. Enhancements in speaker detection and some advanced features could also improve the overall experience Review collected by and hosted on G2.com.