{"slug": "personaplex", "name": "PersonaPlex", "category": "ai-models", "category_name": "AI models", "description": "", "recommenders": 3, "commenters": 9, "warners": 0, "score": 3.901, "rising": -0.98, "first": "2026-01-23", "latest": "2026-03-05", "why": "the NVIDIA Personaplex demos [2] are very impressive in this regard", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 3, "neutral": 4, "compare": 2, "warn": 1}, "spark": [0, 0, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "the NVIDIA Personaplex demos [2] are very impressive in this regard", "kind": "", "icon": false, "domain": "nvidia.com", "github_repo": null, "official_url": "", "pricing": "", "aliases": ["PersonaPlex"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 755605856, "item_id": 47225366, "url": "https://news.ycombinator.com/item?id=47225366", "by": "nicktikhonov", "date": "2026-03-02", "version": null, "stance": "recommend", "quote": "you'll enjoy the new stuff coming out from nvidia", "excerpt": "If you're of that opinion, you'll enjoy the new stuff coming out from nvidia: https://research.nvidia.com/labs/adlr/personaplex/", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: I built a sub-500ms latency voice agent from scratch"}, {"id": 755603746, "item_id": 47225234, "url": "https://news.ycombinator.com/item?id=47225234", "by": "nicktikhonov", "date": "2026-03-02", "version": null, "stance": "recommend", "quote": "try PersonaPlex, Nvidia's new model that would completely replace this architecture", "excerpt": "I didn't try Soniox, but I made a note to check it out! I chose Flux because I was already using Deepgram for STT and just happened to discover it when I was doing research. It would definitely be a good follow-up to try out all the different endpointing solutions to see what would shave off additional latency and feel most natural. Another good follow-up would be to try PersonaPlex, Nvidia's new model that would com", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: I built a sub-500ms latency voice agent from scratch"}, {"id": 748645296, "item_id": 46790331, "url": "https://news.ycombinator.com/item?id=46790331", "by": "jonsoft", "date": "2026-01-27", "version": null, "stance": "recommend", "quote": "the NVIDIA Personaplex demos [2] are very impressive in this regard", "excerpt": "I asked the Spanish tutor if he/it was familiar with the terms seseo[0] and ceceo[1] and he said it wasn't, which surprised me. Ideally it would be possible to choose which Spanish dialect to practise as mainland Spain pronunciation is very different to Latin America. In general it didn't convince me it was really hearing how I was pronouncing words, an important part of learning a language. I would say the tutor is ", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: LemonSlice \u2013 Give your voice agents a face"}, {"id": 747814736, "item_id": 46738421, "url": "https://news.ycombinator.com/item?id=46738421", "by": "mpaepper", "date": "2026-01-23", "version": null, "stance": "recommend", "quote": "You should look into the new Nvidia model", "excerpt": "You should look into the new Nvidia model: https://research.nvidia.com/labs/adlr/personaplex/ It has dual channel input / output and a very permissible license", "context": "", "use_case": "speech-to-speech", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 748706498, "item_id": 46794156, "url": "https://news.ycombinator.com/item?id=46794156", "by": "armcat", "date": "2026-01-28", "version": null, "stance": "warn", "quote": "couldn't get even PersonaPlex to run properly", "excerpt": "This is so awesome, well done LemonSlice team! Super interesting on the ASR->LLM->TTS pipeline, and I agree, you can make it super fast (I did something myself as a 2-hour hobby project: https://github.com/acatovic/ova). I've been following full-duplex models as well and so far couldn't get even PersonaPlex to run properly (without choppiness/latency), but have you peeps tried Sesame, e.g. https://app.sesame.com/? I ", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: LemonSlice \u2013 Give your voice agents a face"}]}], "who": [{"by": "cbrews", "context": "audio background", "url": "https://news.ycombinator.com/item?id=46739544", "date": "2026-01-23"}], "tags": [], "use_cases": ["speech-to-speech model", "full-duplex speech model", "speech-to-speech"], "compared": [{"slug": "moshi", "name": "Moshi", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [{"id": 756172752, "item_id": 47260797, "url": "https://news.ycombinator.com/item?id=47260797", "by": "taf2", "date": "2026-03-05", "version": null, "stance": "recommend", "quote": "it works well for me to toggle lights on and off", "excerpt": "I forked and added tool calling by running another llm in parallel to infer when to call tools it works well for me to toggle lights on and off. Code updates here https://github.com/taf2/personaplex", "context": "", "use_case": "speech-to-speech with tool calling", "switched_from": "", "story": "Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex Speech-to-Speech in Swift"}], "threads": [{"url": "https://news.ycombinator.com/item?id=47258801", "title": "Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex Speech-to-Speech in Swift", "mentions": 4}, {"url": "https://news.ycombinator.com/item?id=47224295", "title": "Show HN: I built a sub-500ms latency voice agent from scratch", "mentions": 3}, {"url": "https://news.ycombinator.com/item?id=46783600", "title": "Show HN: LemonSlice \u2013 Give your voice agents a face", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=46731068", "title": "Ask HN: What's the current best local/open speech-to-speech setup?", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=47219668", "title": "Ask HN: Who is hiring? (March 2026)", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {"recommend": 2, "neutral": 1, "warn": 1}, "2026-02": {}, "2026-03": {"neutral": 3, "recommend": 1, "compare": 2}, "2026-04": {}, "2026-05": {}, "2026-06": {}, "2026-07": {}, "2026-08": {}, "2026-09": {}, "2026-10": {}}, "versions": [], "version_unnamed": 9, "family_ref": null, "runs_on": null, "lines": [], "quotes": [{"id": 756234336, "item_id": 47264646, "url": "https://news.ycombinator.com/item?id=47264646", "by": "reactordev", "date": "2026-03-05", "version": null, "stance": "neutral", "quote": "", "excerpt": "https://gist.github.com/gabereiser/cd8c67262717afd2539dc9c3d00d2c2f", "context": "", "use_case": "", "switched_from": "", "story": "Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex Speech-to-Speech in Swift"}, {"id": 756200624, "item_id": 47262539, "url": "https://news.ycombinator.com/item?id=47262539", "by": "scotty79", "date": "2026-03-05", "version": null, "stance": "neutral", "quote": "This is interactive:", "excerpt": "This is interactive: https://github.com/NVIDIA/personaplex", "context": "", "use_case": "Speech-to-speech model", "switched_from": "", "story": "Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex Speech-to-Speech in Swift"}, {"id": 756152160, "item_id": 47259510, "url": "https://news.ycombinator.com/item?id=47259510", "by": "armcat", "date": "2026-03-05", "version": null, "stance": "compare", "quote": "this full duplex architecture is still a bit off in terms of giving you good accuracy/performance", "excerpt": "I really like this, and have actually tried (unsuccessfully) to get PersonaPlex to run on my blackwell device - I will try this on Mac now as well. There are a few caveats here, for those of you venturing in this, since I've spent considerable time looking at these voice agents. First is that a VAD->ASR->LLM->TTS pipeline can still feel real-time with sub-second RTT. For example, see my project https://github.com/aca", "context": "", "use_case": "Full-duplex speech model", "switched_from": "", "story": "Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex Speech-to-Speech in Swift"}, {"id": 755610689, "item_id": 47225668, "url": "https://news.ycombinator.com/item?id=47225668", "by": "woodson", "date": "2026-03-02", "version": null, "stance": "compare", "quote": "Since Personaplex is just a finetuned Moshi model", "excerpt": "You mean Moshi (https://github.com/kyutai-labs/moshi)? Since Personaplex is just a finetuned Moshi model.", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: I built a sub-500ms latency voice agent from scratch"}, {"id": 755605856, "item_id": 47225366, "url": "https://news.ycombinator.com/item?id=47225366", "by": "nicktikhonov", "date": "2026-03-02", "version": null, "stance": "recommend", "quote": "you'll enjoy the new stuff coming out from nvidia", "excerpt": "If you're of that opinion, you'll enjoy the new stuff coming out from nvidia: https://research.nvidia.com/labs/adlr/personaplex/", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: I built a sub-500ms latency voice agent from scratch"}, {"id": 755603746, "item_id": 47225234, "url": "https://news.ycombinator.com/item?id=47225234", "by": "nicktikhonov", "date": "2026-03-02", "version": null, "stance": "recommend", "quote": "try PersonaPlex, Nvidia's new model that would completely replace this architecture", "excerpt": "I didn't try Soniox, but I made a note to check it out! I chose Flux because I was already using Deepgram for STT and just happened to discover it when I was doing research. It would definitely be a good follow-up to try out all the different endpointing solutions to see what would shave off additional latency and feel most natural. Another good follow-up would be to try PersonaPlex, Nvidia's new model that would com", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: I built a sub-500ms latency voice agent from scratch"}, {"id": 755550224, "item_id": 47221889, "url": "https://news.ycombinator.com/item?id=47221889", "by": "navanchauhan", "date": "2026-03-02", "version": null, "stance": "neutral", "quote": "it uses a duplex architecture", "excerpt": "Not affiliated with Sesame, but this is what the realtime models are trying to solve. If you look at NVIDIA\u2019s PersonaPlex release [0], it uses a duplex architecture. It\u2019s based on Moshi [1], which aims to address this problem by allowing the model to listen and generate audio at the same time. [0] https://github.com/NVIDIA/personaplex [1] https://arxiv.org/abs/2410.00037", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Who is hiring? (March 2026)"}, {"id": 748706498, "item_id": 46794156, "url": "https://news.ycombinator.com/item?id=46794156", "by": "armcat", "date": "2026-01-28", "version": null, "stance": "warn", "quote": "couldn't get even PersonaPlex to run properly", "excerpt": "This is so awesome, well done LemonSlice team! Super interesting on the ASR->LLM->TTS pipeline, and I agree, you can make it super fast (I did something myself as a 2-hour hobby project: https://github.com/acatovic/ova). I've been following full-duplex models as well and so far couldn't get even PersonaPlex to run properly (without choppiness/latency), but have you peeps tried Sesame, e.g. https://app.sesame.com/? I ", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: LemonSlice \u2013 Give your voice agents a face"}, {"id": 748645296, "item_id": 46790331, "url": "https://news.ycombinator.com/item?id=46790331", "by": "jonsoft", "date": "2026-01-27", "version": null, "stance": "recommend", "quote": "the NVIDIA Personaplex demos [2] are very impressive in this regard", "excerpt": "I asked the Spanish tutor if he/it was familiar with the terms seseo[0] and ceceo[1] and he said it wasn't, which surprised me. Ideally it would be possible to choose which Spanish dialect to practise as mainland Spain pronunciation is very different to Latin America. In general it didn't convince me it was really hearing how I was pronouncing words, an important part of learning a language. I would say the tutor is ", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: LemonSlice \u2013 Give your voice agents a face"}, {"id": 747832706, "item_id": 46739544, "url": "https://news.ycombinator.com/item?id=46739544", "by": "cbrews", "date": "2026-01-23", "version": null, "stance": "neutral", "quote": "I'm going to put this on my list to play around with.", "excerpt": "Thanks for sharing this! I'm going to put this on my list to play around with. I'm not really an expert in this tech, I come from the audio background, but recently was playing around with streaming Speech-to-Text (using Whisper) / Text-to-Speech (using Kokoro at the time) on a local machine. The most challenging part in my build was tuning the inference batch sizing here. I was able to get it working well for Speech", "context": "audio background", "use_case": "speech-to-speech model", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 747814736, "item_id": 46738421, "url": "https://news.ycombinator.com/item?id=46738421", "by": "mpaepper", "date": "2026-01-23", "version": null, "stance": "recommend", "quote": "You should look into the new Nvidia model", "excerpt": "You should look into the new Nvidia model: https://research.nvidia.com/labs/adlr/personaplex/ It has dual channel input / output and a very permissible license", "context": "", "use_case": "speech-to-speech", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}]}