I find it very much useful in my workplace meetings where I need to frequently interact with foreign clients and when they are now able to convey the requirements to me easily in their native language, and that was transcribed in to english in email and vice versa for the client Review collected by and hosted on G2.com.

Obviously it’s evolving. Sometimes there are anomalies, since there are many variable dependencies on accent, grammar, and the language itself, but Google is heading in the right direction. Review collected by and hosted on G2.com.

The speech is very clear, and the dictation is quick and responsive. The support for multiple accents makes it easier to communicate without having to change my tone. Review collected by and hosted on G2.com.

The cost goes up when you use high-volume audio. Background noise can also reduce performance, and you can’t use it offline—you need to stay connected to the internet at all times. Review collected by and hosted on G2.com.

I’ve been using this product for the past several years. Since it become available. I would say it’s so much better than your typical iPhone speech to text interpreter.

I am a very busy health insurance broker so I do text a lot to my clients and email, and this tool just makes it so much easier, and I very rarely have to make any edits. Review collected by and hosted on G2.com.

I don’t have anything specifically negative to say about this product, but sometimes it misses a word which is normal. You have to kind of talk like a robot to make it perfect. Review collected by and hosted on G2.com.

The text-to-speech output is accurate and impressively specific when it comes to identifying certain keywords. It also seems to remember my vocabulary from previous conversations, which helps create a more accurate transcription of speech. Review collected by and hosted on G2.com.

The speed of recognising multiple languages differs. English, as the primary language, is identified more accurately than native languages like Hindi or Marathi, especially when they’re spoken with slang mixed in, which can be misinterpreted or translated incorrectly. Review collected by and hosted on G2.com.

In an era of hands-free devices, this makes it much easier to communicate while I’m on the go. My car even allows me to utilize Google Cloud Speech to Text while driving, which helps me stay focused on the road and keeps safety in mind for both myself and my fellow travelers. Review collected by and hosted on G2.com.

There are still some things AI struggles with, like accents and intonation. That said, I believe these issues will be addressed in future updates as the AI continues to learn and improve. Review collected by and hosted on G2.com.

It's pretty accurate. I tried a few STT solutions and GCP's model handles background noises way better than I expected. It supports multiple languages and dialects so that's really a plus point when your aim is to build something for diverse users. Integrations with GCP ecosystem are pretty good. Really love the UX and the performance too. Onboarding through GCP console was pretty smooth for me. Review collected by and hosted on G2.com.

It's the pricing for me. When you use it casually, or you're not careful about tracking your usage, it can really cost you. It adds up faster than expected. Review collected by and hosted on G2.com.

Google Cloud Speech-to-Text is an extraordinary tool for anyone who needs to integrate advanced voice functionalities within a software infrastructure. If I had to highlight the aspects I find most interesting and truly effective, I would focus on these elements.

The first is the power of the Chirp model and its evolution into Chirp 3. What strikes me the most is the approach based on foundation models, instead of the classic supervised systems trained language by language. Chirp (now at version Chirp 3) maintains remarkable accuracy even with particular accents, dialects, or languages historically less covered by traditional models. Moreover, it handles natural and spontaneous speech very well, which often challenges more "rigid" transcribers.

Another strong point is the noise robustness. In real life, people do not record in a professional studio: they speak on the street, in cars, or in noisy offices. The algorithm's ability to clean the audio track in real-time and extract speech without requiring heavy pre-processing of the signal represents an enormous added value, especially from the perspective of those who develop and need to integrate the solution in uncontrolled environments. Review collected by and hosted on G2.com.

It is extremely strict regarding input formats (such as sampling rate or encoding). If the audio does not exactly meet the parameters specified in the JSON request, the API does not adapt: it either fails or returns silence. Review collected by and hosted on G2.com.

We use Google Cloud speech to text for generating transcript for our vide feeds for linear TV and CTV aka streaming data. Review collected by and hosted on G2.com.

It’s expensive when you have a large volume of speech-to-text, so we sometimes need to use free models like Whisper. Review collected by and hosted on G2.com.

High Accuracy & Multilingual Support: The system supports over 125 languages and dialects, delivering reliable, context-aware transcriptions even with challenging audio.

Chirp 3 Foundation Model: It utilizes advanced, large-scale models trained on millions of hours of audio, improving accuracy for global, varied speech patterns.

Real-Time & Batch Processing: It handles both streaming (live) and pre-recorded audio efficiently.

Model Adaptation: Users can customize the model with specific vocabulary and phrases to improve recognition of industry-specific jargon.

Intelligent Features: The service includes automated punctuation, speaker diarization (identifying who is speaking), and timestamping Review collected by and hosted on G2.com.

Low Accuracy in Noisy Environments: Uses advanced machine learning and model adaptation to recognize specific, specialized vocabulary and improve transcription quality.

Language Barriers: Supports over 125 languages and variants, enabling global, multi-lingual transcription capabilities. Review collected by and hosted on G2.com.

It’s very accurate at decoding speech into text. Review collected by and hosted on G2.com.

Nothing! The program is very accurate when speaking and recording. Review collected by and hosted on G2.com.