June, 2026

Introduction

This release brings enhancements across Accent Translation, Language Translation, the Sanas Portal, and the Sanas Chrome Extension. The focus areas include improved communication clarity, simplified setup and administration, optimized performance, and a more seamless user experience through smarter automation, enhanced audio controls, and expanded accessibility features.

What’s New

• Inbound Accent Translation: Real-time translation of incoming audio accents to improve comprehension. Inbound Accent Translation can be used independently or alongside Outbound Accent Translation.

• Sanas now automatically applies the optimal voice-isolation mode based on the connected headset:

  • Wired USB headsets → Near-field mode

  • Bluetooth or other devices → General mode

• Acoustic Echo Cancellation (AEC): Automatically eliminates echoed audio during conversations by preventing speaker output from being picked up by the microphone. AEC activates when both Sanas Microphone and Sanas Speaker are used in the same conferencing app. No manual configuration is required.

• Smart Microphone & Speaker Activation: Reduces resource usage by automatically pausing Sanas audio pipelines when no application is connected to the virtual microphone or speaker. Pipelines resume when needed, and model inference only runs during active audio sessions. Behavior is controlled through the Stream Control setting in the Sanas Portal.

• Users can now mute the Sanas Virtual Speaker directly from the operating system's volume mixer.

  • Mute behavior is configured by admins in portal settings.

  • Disabled by default.

  • The speaker automatically unmutes when the setting is turned off.

• Volume Sync Behaviour: Admins can independently control volume adjustment and synchronization for the Sanas Virtual Microphone and Speaker devices, with settings managed through the Sanas Portal. And the available modes:

  • Locked 100% volume — volume is fixed and cannot be adjusted by users.

  • Flexible user-controlled volume — users can freely adjust volume levels.

  • Bidirectional sync — volume and mute state stay in sync with the underlying physical device.

• Language Translation

  • Audio Crossfade: Smoothly lowers the original speaker's audio as translated speech begins, keeping the speaker faintly audible rather than cutting them off abruptly. This reduces perceived latency, maintains speaker awareness, and keeps translated speech clear.

  • Auto Mode: Automatically detects the speaker's language after a short period of speech and updates translation settings accordingly. Once a language is identified, it remains active until Language Translation is turned off and back on to reset detection.

  • Latency Metrics Display: Provides precise, per-utterance latency measurements — useful for debugging, demos, and POC testing. Disabled by default to avoid distraction during normal operation.

• Sanas Chrome Extension

  • SSO (Single Sign-On) Authentication: A new login method that allows users to securely access the Chrome Extension using their organization's identity provider. SSO enhances security through centralized authentication while providing a more seamless login experience.

  • Automatic Sanas Microphone Configuration: Sanas Microphone is now automatically set as the default microphone for web-based conferencing apps and softphones. Users benefit from Accent Translation without any manual setup.

Inbound Accent Translation

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• Inbound Accent Translation (Inbound AT) translates the accent of incoming audio in real time, improving comprehension during conversations.

• Works independently of Outbound AT, allowing users to enable:

  • Inbound AT only

  • Outbound AT only

  • Both Inbound and Outbound AT

  • Neither feature

• How it works?

  • Enable Inbound AT in the Sanas portal by following this path:

    • Sanas Portal → Settings → Configuration → Products → Accent Translation → Accent Translation Access.

      

  • In the Sanas app, toggle Inbound AT under “What You Hear”.

    

  • Ensure Sanas Speaker is selected as the output device in your conferencing application.

Headset-based Model Detection for Accent Translation

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• Headset-Based Model Detection for Accent Translation automatically selects the appropriate voice-isolation mode based on the headset being used.

  • Near-field mode is applied for wired USB headsets.

  • General mode is applied for Bluetooth and other headset types.

• Previously available only for Speech Enhancement, this capability is now extended to Accent Translation.

• How it works?

  • Sanas Portal → Settings → Configuration → Products → Accent Translation → Outgoing Audio → Enhancement Mode → Voice Isolation → Microphone Detection → Auto.

    

  • Override the automatically selected mode for specific headsets through the app’s headset list, if needed.

Acoustic Echo Cancellation (AEC)

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• Acoustic Echo Cancellation (AEC) prevents remote callers from hearing their own voice echoed back during conversations.

• Eliminates echo caused when audio from speakers is picked up by the microphone.

• Particularly beneficial for:

  • Open-back headsets (e.g., Logitech H390)

  • Built-in laptop microphone and speaker setups

• How it works?

  • Automatically activates when both Sanas Microphone and Sanas Speaker are selected in the same conferencing application.

  • No in-app toggle is required.

  • Availability is managed by admins in the Sanas portal:

    • Sanas Portal → Settings → Configuration → Desktop Application → Application Settings → Echo Cancellation.

      

Smart Microphone & Speaker Activation

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• The Smart Microphone and Speaker Activation feature reduces resource usage by pausing Sanas audio pipelines when they are not in use.

  • Both microphone and speaker pipelines automatically pause when no application is connected to the corresponding virtual device.

  • Pipelines automatically resume as soon as a new audio session attaches.

• The most resource-intensive component — model inference — runs only when there is an active audio consumer.

• How it works?

  • Controlled through the Sanas portal.

    • Sanas Portal → Settings → Configuration → Desktop Application → Application Settings → Pause/Resume Microphone Stream.

      

  • Configuration changes take effect on the next heartbeat; no app restart is required.

Speaker Mute Control

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• Virtual Speaker Mute Control allows the Sanas Virtual Speaker to be muted directly from the operating system’s volume mixer.

• Previously, the Sanas Virtual Speaker was force-unmuted across all platforms.

• Mute behavior is now governed by an administrator-controlled Portal policy.

• How it works?

  • Controlled through the Sanas portal.

    • Sanas Portal → Settings → Configuration → Desktop Application → Application Settings → Mute Sanas Speaker.

      

  • Default setting: Disabled.

  • When the setting is turned Off, the SDK automatically unmutes the Sanas Virtual Speaker if it is currently muted.

Volume Sync Behaviour

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• Volume control for the Sanas Virtual Microphone and Speaker is now driven by two independent Sanas Portal configurations — one per channel — each with two flags:

  • Flexible volume control — when off, the Sanas device's volume locks at 100%; any change is reverted. When on, users can adjust freely.

  • Sync with selected speaker — when on, the Sanas device and the underlying physical device share a single, bidirectionally synced volume (and mute state). Requires the adjustment flag to also be on.

• How it works?

  • Configurable in the Sanas portal.

    • Sanas Portal → Settings → Configuration → Desktop Application → Application Settings.

    • Followed by two entitlement objects:

      • Enable Microphone Volume Adjustment (Sync with Selected Microphone OR Flexible Volume Control)

        

      • Enable Speaker Volume Adjustment (Sync with Selected Speaker OR Flexible Volume Control)

        

  • Disabling either parent entitlement defaults both child flags to off (locked at 100%, no sync). Flags are re-evaluated on every heartbeat — no restart required.

Audio Crossfade for Language Translation

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• A highly anticipated addition, the Audio Crossfade feature facilitates a smooth transition from incoming raw speech to a translated audio stream. By utilizing a ducking percentage, the system ensures that the original audio remains subtly audible rather than cutting out completely.

• To visualize how this works, imagine a room containing a Spanish speaker, an English speaker, and a translator. As the Spanish speaker speaks, you initially hear their native speech, followed by the simultaneous voice of the translator.

• The Language Translation (LT) application manages this process through the following steps once translated speech begins:

  • Raw speech begins to fade out based on a user-defined fade-out duration.

  • The fade-out concludes at a specific ducking percentage, which defaults to 20%.

  • If the translated audio becomes silent for two seconds, the raw audio automatically fades back to full volume (100%). This avoids the audio from bouncing.

• This functionality offers several key benefits:

  • Lower Perceived Latency: Users are not left in silence while waiting for the translation to start.

  • Continuous Confirmation: Ongoing raw audio confirms the input speaker is still active.

  • Clarity: The raw audio volume is lowered sufficiently to ensure the translated speech remains prominent.

Note: Extensive testing has informed our default settings, though we encourage users to share their feedback and experiences.

Auto Mode for Language Translation

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• Another added feature on Language Translation in the app is the Auto Mode.

• Users can enable Auto Mode in the app by changing “They Speak” to Auto.

  

• Upon starting a conversation, the server automatically identifies the language after a brief period of speech. It then informs the application of the verified "They Speak" setting, such as Auto(Spanish LATAM), depending on the detected language.

  

• This process modifies the backend pipeline's Language ID just as if it had been manually selected by the user. To return to the initial Auto state (where no specific language is yet identified), Language Translation must be toggled off and then back on.

• To minimize the need for this manual reset, we are currently exploring call-drop-detection enhancements.

Latency Metrics Display for Language Translation

Sanas Desktop App version: 5.6.7 to 5.6.8 and above

• Designed primarily for debugging, the Latency Metrics Display is a valuable tool for pre-sales demonstrations and Proof of Concept (POC) testing. While this feature can be activated through the portal, it remains disabled by default to avoid creating potential distractions during standard operations.

• We frequently engage in discussions regarding our latency figures, which should range from 1 to 5 seconds. However, customers often lack a clear understanding of the specific metrics these values represent, sometimes even attempting to estimate latency manually, which results in inaccurate assessments.

• To address this, the Latency Metrics Display provides precise data by attaching a dedicated section to every utterance card. This section highlights two critical one-way latency measurements that our team stands by:

  • TTFS (Time To First Speech)

  • TALS (Time After Last Speech)

• When enabled, each card will look like this; notice the values at the bottom.

  

• To configure in the Sanas portal for your group, make sure this is enabled on your Feature Management tab.  It should be on by default.

  

• Within the Configuration tab, these options are available.  Display Translation Latency is only for debugging purposes and should generally be turned off.

  

Note: Sanas portal updates are released with the desktop app. As such, these configuration settings may not be visible to your groups yet. Please reach out to our Support Team for validation and further assistance.

Single Sign-On Login Authentication on the Sanas Chrome Extension

Sanas Chrome Extension version: 1.3.2 and above

• Another login method that allows users to access the Sanas Chrome Extension using their existing company credentials through their organization's identity provider (e.g., Microsoft Entra/Azure ID, Okta, or Google Workspace).

• Enhances security by leveraging organizational authentication policies such as Multi-Factor Authentication (MFA), password requirements, and access controls, eliminating the need for separate Sanas credentials.

  

• For the detailed steps on how to leverage this login authentication method, refer to the following articles below:

  • Overview

  • Quickstart Guide

  • User Guide

Sanas Microphone as a Default Audio Input Device

Sanas Chrome Extension version: 1.4.0 and above

• To ensure optimal use, Sanas Microphone has now been configured as your browser’s default audio input device, enabling meeting applications (such as Google Meet Web, Zoom Web, Webex Web, etc.) and web-based softphones to automatically use Sanas without requiring manual microphone selection.

• For instance, we have leverage Google Meet Web to validate this setup:

  

• If a user logs out of the Chrome Extension, the Sanas Microphone remains visible but Accent Translation is disabled until they sign back in, after which functionality is automatically restored (a tab refresh may be required).

• For the complete details on how this is configured, refer to the guide below.

  • Audio Configuration on Sanas Chrome Extension

Support