Skip to main content
AI Voice AgentsNovember 04, 20265 min read

Soniox Releases Soniox STT v6 with Agent Text Input and Echo Filtering for Voice Agents

Soniox has launched the latest version of its speech-to-text API, Soniox STT v6, designed to improve voice agent interactions by transcribing agent speech while hearing callers, recognizing short replies, and filtering out agent noise. This update supports both streaming and asynchronous modes, offering better real-time and batch transcription for sales and customer service teams.

By BrightLaunchIQ, Intelligence Team
Soniox Releases Soniox STT v6 with Agent Text Input and Echo Filtering for Voice Agents editorial illustration
BrightLaunchIQ AI Industry News

Soniox Introduces Soniox STT v6 for Smarter Voice Agents

Soniox, a company specializing in speech-to-text technology, has released Soniox STT v6, the newest version of its API built specifically for voice agents. This updated model can transcribe what a voice agent says in addition to listening to the caller. This dual listening ability helps the system better recognize short replies, such as "yes" or "no," more accurately during conversations.

Why Agent Text Input and Echo Filtering Matter

One of the key improvements in Soniox STT v6 is its capacity to filter out the voice agent's own speech to reduce echo. In many voice interactions, the agent’s own voice can create noise that interferes with transcription accuracy. By filtering out the agent's voice, the API minimizes this problem, helping to produce clearer transcripts.

This feature is particularly useful because it ensures that voice agents can better understand callers without the confusion of overlapping sounds or repeated audio feedback. It also allows the system to capture short replies effectively, which are common in quick back-and-forth sales or customer service conversations.

Streaming and Asynchronous Modes Enhance Flexibility

Soniox STT v6 supports both streaming and asynchronous transcription modes. Streaming mode offers real-time transcription, ideal for live customer service calls or sales pitches where immediate text output is necessary. Asynchronous mode, on the other hand, handles batch transcription where audio files can be processed after the call ends, useful in quality analysis or training purposes.

With these options, organizations can choose how to integrate the technology depending on their needs and resources.

Benefits for Sales and Customer Service Teams

For sales and customer service teams, Soniox STT v6 can improve call accuracy and provide faster, cleaner texts during or after calls. This results in better record-keeping, allows quick response analysis, and enhances the overall customer experience by reducing misunderstandings.

Small businesses and call centers using voice agents can especially benefit by making conversations more natural and productive. Clearer transcripts help agents follow up on conversation details and offer more personalized service.

What to Watch Next

As voice agents become increasingly common in business settings, technologies like Soniox STT v6 that improve transcription accuracy and real-time understanding will be important to watch. Future updates could focus on expanding language support, improving accuracy in noisy environments, or adding further integration features for popular customer relationship management systems.

Organizations interested in upgrading their voice agent capabilities should keep an eye on how Soniox continues to develop its technology to support smarter and more responsive AI-driven communications.

Keep reading

BrightLaunchIQ Intelligence Team