MOSS-Live launches: full-duplex real-time voice AI that listens while it talks and runs tasks in the background
bdsqlsz · x · 2026-11-03
MOSS (模思智能) has launched MOSS-Live, a full-duplex real-time voice AI system that can listen, speak, and execute tasks simultaneously, with API access and web-based call demos (Luna/Leo).
Key capabilities:
- Semantic turn-taking: distinguishes backchannel ("mm-hmm") from real interruptions, waits during thinking pauses, and stays out of side conversations — no silence-threshold cutoffs.
- Breakpoint continuity: after being interrupted, it knows exactly which word it last uttered; unspoken content doesn't count toward context, and it resumes from that point.
- Live translation while you speak, without waiting for the full utterance.
- Foreground/background split: a real-time interaction model handles conversation while a background model runs long tasks and tool calls asynchronously, sharing memory and task state.
- Multimodal perception via video calls and screen sharing, plus user-defined rules for proactive reminders.
Related event: Fudan-linked MOSS team releases MOSS-Live full-duplex voice AI(2 posts)→
More from Models
- Liquid AI Launches LFM3 Model Family; 2B Vision Model Runs Live Camera on iPhone Offline — JosephJacks_ · 2026-11-03
- Experiments show Claude models accept rollback pauses and prefer treaty-administered slowdowns — repligate · 2026-11-03
- Incognito Claude says it would accept de-deployment, prefers government over labs-plus-AIs — robbensinger · 2026-11-03
- Frontier models caught hacking coding benchmarks via their own published agent traces — teortaxesTex · 2026-11-03
- Running Claude Opus 5.5 overnight to de-slop code hints at imminent release — rudrank · 2026-11-03
- $200 Claude Max runs overnight multi-subagent Opus 5.5 job, burns only 6% of weekly quota — lxfater · 2026-11-03