Speechmatics Reviews

The ASR solution is provided in a containerized form. The API is easy to use. The collaboration with the contacts at Speechmatics is good. Pricing is fair (although now under pressure from new offers).

In the German language model (CPU and GPU), we have not been able to measure any noticeable improvement regarding WER in the latest versions. The language models of other manufacturers are sometimes better for German.

We primarily use the Speechmatics SR engine for transcription. A good or accurate transcription with good speaker separation is essential for further processing steps that we offer with our software. The quality of the transcription ultimately determines the satisfaction of the end customers. Although we would also like to have summaries, these are not yet available in a containerized form.

I really appreciate Speechmatics for its really good below word error rate compared to some of the bigger vendors. They've clearly focused on providing high accuracy and good quality speech-to-text for general-purpose use, especially in languages that are generally hard for other vendors. They also excel in more niche languages where there’s probably minimal data. Their ability to handle a variety of dialects and accents with high accuracy has been really critical for us.

When I initially started working with Speechmatics, they had one of the highest latency levels, which was important for real-time voice. However, this is not an issue anymore.

I use Speechmatics to transcribe callers and customer requests efficiently, enabling AI voice bots to support customers. It's critical for providing accurate speech-to-text, especially in niche languages with high accuracy in various dialects and accents.

Speechmatics made it very simple to set up AI transcription for our product. As with any integration there where snags, but nothing we couldn't figure out. Overall we're very pleased with the results we are achieving, and very pleased with the help we've gotten from their sales and support departments.

I love that they provide open source packages for many of the mechanics we need to implement in relation to transcribing audio to text, however we do have some bugs still related to these that we haven't been able to figure out yet. So far we can live with the issues we've experienced.

Speechmatics provides audio transcription in multiple languages so that we can deliver full AI features to our customers.

.latency and accuracy for Live for the captioning

Multichannel use case and dubbing not supported

Used for live captioning &

The high accuracy and quality of live transcription metadata (including speaker detection) has made it possible for DR to develop and deploy a real-time subtitling solution for news, sports and political debates. This is used for national broadcasts in Denmark for approximately 6 hours daily.

There is nothing I can think of. We have always had good support and input towards future features.

Real-time transcription is used for generating subtitles during live broadcasts. File-based transcription will be used increasingly for preparation of subtitles in pre-recorded programmes.

It delivers highly accurate transcription with very low latency, so the text appears quickly and stays reliable.

It didn't support multichannel audio for live streams, so I had to connect each channel separately. I would love to see them add voice cloning and dubbing features like the ones on ElevenLabs.

Our use case focused on live captioning for broadcasters. Speechmatics generated highly accurate text from the live audio stream, which we converted into captions and delivered on air alongside the video.

High marks for the accuracy of the translations and transcriptions, along with the API’s high throughput.

The SDK is quite decent and workable, but it didn’t feel as mature as a competitor’s.

Accurate translation and unmatched ASR.

These guys are incredible. These guys are incredible.

diarization needs to get better diarization needs to get better

Solving keep the cost low when it comes to transcription for various cases, e.g.

We compared a few providers and found Speechmatics' technology to be more accurate.

I wish the free/trial plan was a bit more robust but their sales team is really great to work with and offered some flexibility as we got up and running.

Accurately identifying live speakers and transcribing appropriately.

I like it because we use it a the transcription layer to our research tool. It has several languages integrated to it so whenever a user uploads a lecture that has English and French, the system can easily recognize it. Our main usage is the speech to text and its perfect. Lastly, the ease of integration for the API.

Coming from Africa, we have diverse local languages and this are not recognized in speechmatics. That is one downside we have experienced.

Speechmatics API is allowing our platform(www.databridge.co.ke) to have transcription capabilities so that our users can transcribe their video and audio for research purposes on the platform.