What's new in Connect
Every release, every fix, every improvement — in one place. The most recent update is always at the top.
Replaced HTTP polling with a WebSocket pipeline for sub-200 ms audio round-trips.
Reset Audio Devices button in profile — one click to restore any broken routing state.
Virtual Cable installer now creates Connect Mic and Connect Speaker devices automatically.
Audio service restart after install — devices are ready without a system reboot on most machines.
Speaker Labeling — transcripts now track who said what in group calls.
Language Hints — tell Connect which languages to expect for higher accuracy on multilingual calls.
Emotion Transfer (Pro) — translated voice now carries your original tone and energy.
Noise Suppression toggle in settings to filter background audio before processing.
Streaming Mode latency reduced to ~480 ms average on stable connections.
Voiceprint — record or upload a clip so Connect keeps your translated voice close to your identity.
Voice Variants — create separate voice profiles for different languages and meeting types.
Pronunciation Dictionary — add custom word pronunciations for names, acronyms and brands.
Conversation Context profiles — tag meetings with topics (medical, legal, sales…) to boost terminology accuracy.
How I Sound? — test your translated voice privately before a real call.
Portal redesign with unified Voice Library, Context and Dictionary views.
Bidirectional Mode — translate both sides of the conversation simultaneously.
Instant Mode — cleaner short-burst translations after natural pauses, alongside Streaming Mode.
Linux support — Connect now runs on Ubuntu 20.04+ and Debian-based distributions.
40+ language pairs available with full voice preservation across all of them.
CONNECT-ON state correctly restored after audio device changes mid-session.
First public release. Real-time AI voice translation under 500 ms on Zoom, Google Meet, Teams, Slack and any app that uses a microphone.
Virtual Audio Cable (Windows) — Connect Mic and Connect Speaker devices for seamless routing.
Unidirectional mode — translate your outbound voice without touching incoming audio.
Free plan with 10,000 credits/month. Standard and Pro plans for heavier use.
Zero audio stored — all processing is live and ephemeral, nothing is saved.
What's coming next
These are the features currently in progress or planned for the next releases.