Microsoft VibeVoice: Open-Source Frontier Voice AI

(github.com)

99 points | by tosh 2 hours ago

15 comments

  • maxloh 57 minutes ago
    I think we should stop calling this type of models open source. They are indeed "open weight." The training code is proprietary and never revealed.

    https://github.com/microsoft/VibeVoice/issues/102

    • jcmfernandes 9 minutes ago
      Indeed. We now live in a world where freeware is named open source. We are very sorry, Stallman.
    • JumpCrisscross 47 minutes ago
      > we should stop calling this type of model open source. They are indeed "open weight”

      This ship has sailed. It’s now in the same category as hacker/cracker and the pronunciation of GIF.

      • andy_ppp 46 minutes ago
        I think you mean GIF.
      • giancarlostoro 34 minutes ago
        It's the same as GIS, you wouldn't say jizz now would you?
        • dijksterhuis 0 minutes ago
          i am absolutely going to from now on
        • kevin_thibedeau 13 minutes ago
          The developer of the format declared the pronunciation 30+ years ago. It has always been jif.
        • pardon_me 6 minutes ago
          How do you pronounce giraffe?
          • parineum 2 minutes ago
            How do you pronounce gift?
        • DoctorOW 19 minutes ago
          I absolutely do, every single time it comes up.
        • notabotiswear 27 minutes ago
          I take it that you haven’t met the Arcgees people…
      • WarmWash 20 minutes ago
        And "hallucination" which should have been "delusion".

        Way early on (spring 2023) people tried to stop it, but no luck.

    • giancarlostoro 35 minutes ago
      I mean, you have "AI" which means just about anything in marketing speak, "Agentic" is kind of becoming similar, hopefully they don't goof that one too badly, would be nice to know what you are trying to sell me. Used to be "Cloud" meant storage not just hosting (I guess it still does).

      Then there's "Smart" in front of Car, Phone, TV, and so on... Meaning different things.

      I do think "Open Weight" should be more commonly used. There's definitely communities that spring up that build the training infrastructure and inference infrastructure around open models on the other hand.

    • notabotiswear 24 minutes ago
      Openwashing is the new greenwashing, which, coincidently, seems to have gone out of fashion a few hundred datacentres ago.
      • dist-epoch 16 minutes ago
        it was replaced with abundancewashing
  • steinvakt2 1 hour ago
    This is not a new model. Also, it hallucinates a lot. Also, it's very heavy and slow in inference. It's also bad in multilingual.

    Edit: I'm talking purely about speech to text (STT). Not sure about the other things this can do.

  • frangonf 13 minutes ago
    I took a look into local options for ASR and diarization some months ago, I missed that VibeVoice now has this feature.

    My conclusions back then (which only came from a shallow research on the topic and 0 real experience mind you) was that Whisper + Pyannote was the "stable" approach.

    Have the VibeVoice, Voxtral, Qwen or the Nemo solutions caught up in segmentation and speaker recognition?

  • Void_ 58 minutes ago
    I the past month or so, I added 2 models to my app Whisper Memos (https://whispermemos.com):

    - Cohere Transcribe (self hosted)

    - Grok Speech To Text (they provide an API, only $0.10/hr!)

    They are both excellent. I'm not sure about this one. Would you like to see it in a consumer speech to text app?

    • Barbing 9 minutes ago
      Does Cohere work with longer transcripts? Do you have to do some magic to merge recordings over 35 seconds long?
    • olejorgenb 48 minutes ago
      I've had good experiences with the Mistral Voxtral models (I've used the API, but some of the model-variants are open weight)
    • 2ndorderthought 47 minutes ago
      Have you tried qwen?
    • SecretDreams 45 minutes ago
      Any non-Musk alternatives that are comparable in quality and cost?
      • jayphen 6 minutes ago
        Voxtral competes on price ($0.003/min) and quality. Speechmatics has best in class accuracy but is a bit more expensive ($0.004/min)
      • Void_ 38 minutes ago
        Our default is still OpenAI Whisper. Grok is just a choice for users who might prefer it.
  • aqme28 41 minutes ago
    Interesting to see "vibe" enshrined by the likes of Microsoft as an AI product word.
    • accrual 31 minutes ago
      Especially when "vibe coded" can have a negative connotation meaning quickly put together without understanding.
      • Barbing 7 minutes ago
        I’m just surprised they put the name of the e-waste slop company in their product
    • altmanaltman 18 minutes ago
      Which makes it even more weird they get offended when people use Mircoslop. They are the ones leaning into the marketing
  • embedding-shape 1 hour ago
    Isn't this project the one Microsoft published but then soon after pulled it for security/safety reasons? What has changed since then?
    • infecto 21 minutes ago
      It’s confusing (at least for me) because the project covers a number of things including what you are mentioning.
    • 542458 1 hour ago
      Look at the "News" section in the readme - The original TTS model is gone from this repo (you can still find it other places), but the SST/ASR, long form TTS, and streaming TTS models are newer.
  • CubsFan1060 1 hour ago
    Great post last night from Simon: https://simonwillison.net/2026/Apr/27/vibevoice/
    • 542458 1 hour ago
      Note that this just covers the Speech-to-Text/Speech-Recognition aspect (a-la whisper), there's also models for long-form Text-To-Speech and steaming Text-To-Speech.
    • JumpCrisscross 44 minutes ago
      “VibeVoice can only handle up to an hour of audio”

      Why?

  • ryukoposting 21 minutes ago
    Holy moly, a Microsoft AI product that isn't named Copilot!
    • DoctorOW 18 minutes ago
      Missed opportunity to call it Vopilot
  • podgietaru 1 hour ago
    So we've really just settled on Vibe as the verb for AI then?
    • giarc 1 hour ago
      I'd be willing to bet it will be "Word of the Year" for 2026. Merriam-Webster had 'slop' for 2025, and 'polarization' for 2024. Is there a prediction market for this?
      • internet_points 47 minutes ago
        it'll probably be something we're not even talking about yet - we still have 7 months in which to make the world even worse
    • pryanshu89 1 hour ago
      Why use precise technical language when you can just vibe with your AI system?
  • JumpCrisscross 41 minutes ago
    What’s the current state of the art, for each of training locally and in the cloud, for learning my voice?
    • yreg 5 minutes ago
      Locally maybe https://voicebox.sh/

      Elevenlabs in the cloud.

    • chrsw 31 minutes ago
      Local? No idea. Cloud? Eleven Labs, probably. But it's described as "cloning" not "training". Not sure what the distinction is or why it matters if the end result is you can to generate any TTS that sounds like you. There might very well be an important one, I just don't know it.
  • Anonyneko 1 hour ago
    You have selected Microsoft Sam as the computer's default voice.
    • accrual 29 minutes ago
      My friends and I had fun in the computer lab with Microsoft Sam, inputting long strings of characters to create funny sound effects. Sususususususu.
  • pluc 51 minutes ago
    Interesting story about this repo/product/author by cybersecurity researcher Kevin Beaumont: https://cyberplace.social/@GossiTheDog/116454846703138243
  • BlastBash192 39 minutes ago
    Maybe Microsoft’s real strength was never making the best model, it was knowing you don’t need to, as long as you own the platform everyone builds on.
  • mistic92 50 minutes ago
    For me its giving me very poor results
  • walthamstow 1 hour ago
    Seems quite heavy for a STT model, Parakeet and Whisper are much smaller and perform great for quick dictation and transcription of longer files. I guess that's due to additional accuracy and speaker diarisation?

    The TTS example clip in the repo of 'spontaneous singing' is creepy as fuck