{
  "id": 20509378,
  "title": "I replaced ElevenLabs with a free open-source tool that runs entirely on my machine, without sending anything to the cloud",
  "url": "https://urgent.news/2026/11/17/i-replaced-elevenlabs-with-a-free-open-source-tool-that-runs-entirely",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-11-17T22:30:21.000Z",
  "source": {
    "name": "XDA Developers",
    "slug": "xda-developers",
    "url": "https://www.xda-developers.com/replaced-elevenlabs-with-a-free-open-source-tool-that-runs-entirely-on-my-machine/"
  },
  "original_language": "en",
  "account": "VoiceStudio, an open-source alternative to ElevenLabs, can run entirely on your personal machine without sending any data to the cloud. This tool offers voice cloning, video dubbing, and transcription services, all operating independently. To test its capabilities, a desktop PC with a powerful GPU was used.\n\nUpon installation, VoiceStudio initially struggled to function, requiring ample time to load its Python environment. However, once it was up and running, it demonstrated impressive results. The tool was able to clone a user's voice and dub a German documentary clip into English. Most notably, the clone accurately replicated the user's pacing, pitch, and accent, as well as their breath sounds, which is crucial for achieving a natural-sounding result.\n\nThe tool's performance is notably faster than traditional methods. For example, it generated about 20 seconds of audio in just 2.3 seconds on an RTX 5090 GPU, achieving speeds approximately eight times faster than real-time. However, the tool is not flawless. It occasionally mispronounces words and struggles with certain sound endings. Additionally, there are instances where the built-in voices sound metallic or obviously synthetic.\n\nTo improve the dubbing process, users need to create a voice profile using their own audio. This results in clearer dubbing, though some hissing sounds may still be present. The translation feature, while generally accurate, requires proofreading. Despite this, the overall performance is impressive, particularly when the tool operates offline without the need for an internet connection.",
  "summary": "This offline-capable voice cloner only needed 30 seconds of audio to sound like me",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}