Deepgram Reviews

Deepgram provides TTS and STT all at a place, the transcription works pretty well, With high accuracy. Api documentation is sok good it's easier to integrate Deepgram in the voice agents.

Pricing can get expensive at scale, While it's TTS is strongest part of it, STT is on the weaker side as it doesn't have many languages support so you might wanna integrate other TTS software to it

Deepgram is best at transcribing and I've used it for making the ai agents with a LLM and a different TTS.

I like how easy it is to use, and the latency is low enough for real-time applications like Voice AI agents.

Only a few Indic languages are supported for STT, which limits its usefulness for my needs.

It solves the STT part for us and gives us low-latency speech-to-text in our voice AI stack.

I like Deepgram's semantic turn detection as it's much better compared to its competitors. We have an internal benchmark for transcription accuracy and noise background cancellation, and Deepgram does the best in that. The initial setup seemed relatively easy, which was great.

Yeah. Unfortunately, right now, we're forced to use Assembly and Azure for multilingual stuff. For multilingual turn like, turn detection based speech to text, and they're much more expensive. Well, not assembly, but Azure is much more expensive, and assembly is horribly inaccurate.

I switched to Deepgram for better semantic turn detection and transcription accuracy. It excels in canceling noise and maintaining session flow, which is crucial for student engagement and preventing churn.

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.

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.

I use Deepgram to create subtitles for YouTube videos, which sync perfectly with the voiceover and automate my work.

I like Deepgram for its transcription accuracy, which really helps me. I also appreciate the API documentation and model efficiency. The API documentation made it easy for me to set up an accurate speech-to-text service for my voice agent, and the model helps with accuracy.

I was using this for TTS services but sentiment tags were not to the mark sometimes, so I think TTS services can be improved.

I use Deepgram for speech-to-text, and it enhances my transcription accuracy. The API documentation was crucial in setting up an accurate STT service for my voice agent, and the model efficiency boosts accuracy.

I like Deepgram's ability to turn spoken conversation into structured and usable data for meetings, which makes it incredibly helpful. The real-time feel is impressive, as it genuinely feels live. I also appreciate that it is not a laggy software and it's specifically built for conversation, with good quality. The initial setup was very easy, and they provide instructions that are really useful.

I didn't notice anything that is not working but can improve UI

I use Deepgram to convert spoken conversation into structured and usable data for meetings.

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.

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.

I use Deepgram for my voice agent project, appreciating its low latency streaming and reliable transcripts for text-to-speech. It solves voice AI challenges by making real-time calls smooth and natural. It's easy to set up, cost-effective compared to others, and handles live calls by instantly converting conversations into text.

I am generally using Deepgram for its voice-based feature, which is incredibly helpful as I am making an AI movie and rely on the text-to-speech feature. The voice quality is what I like the most, as it is cost-effective and provides a more natural human-like voice. The affordability is a big reason why we're shifting from Eleven Labs to Deepgram. Additionally, the setup is easy, and the documentation is very clear.

I think emotional control to be there in voice to express the voice more naturally

I use Deepgram for its voice features, which are cost-effective and sound more natural and human-like, crucial for my AI movie project.

I use Deepgram for converting audio to text format, and it helps convert very accurately. It's good, fast, and accurate, handling every process smoothly. The real-time processing helps me without more waiting, and I enjoy this tool very much.

Till now, I didn't face any issues, so for me, nothing to be improved, but the setup process should be made simple.

I use Deepgram for converting audio to text accurately and quickly. It solves the problem of manual transcription by processing in real-time, which saves me a lot of time.

I use Deepgram for STT services and I felt it has so much accuracy, which is really important for transcription. I like its well-documented structure and models. The API docs are particularly helpful for understanding and setting up the plan. It also helps to know which model has what capabilities, thanks to the Deepgram dashboard and playground for testing models and services. The initial setup was good and all credit goes to the documentation.

I felt a little bit of issue with the TTS service, as it felt more stuck sometimes instead of better transcribing. I think the TTS service can be improved by making its latency better and adding sentiments.

I use Deepgram for transcription with high accuracy. Its well-documented structure and API docs make setting up models easy. The dashboard and playground help in testing models and understanding capabilities.