[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:pocket-tts":3},"\u003Ch1>Pocket TTS\u003C\u002Fh1>\n\u003Cimg width=\"1446\" height=\"622\" alt=\"pocket-tts-logo-v2-transparent\" src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F637b5ed6-831f-4023-9b4c-741be21ab238\" \u002F>\u003Cp>A lightweight text-to-speech (TTS) application designed to run efficiently on CPUs.\nForget about the hassle of using GPUs and web APIs serving TTS models. With Kyutai's Pocket TTS, generating audio is just a pip install and a function call away.\u003C\u002Fp>\n\u003Cp>Supports Python 3.10, 3.11, 3.12, 3.13 and 3.14. Requires PyTorch 2.5+. Does not require the gpu version of PyTorch.\u003C\u002Fp>\n\u003Cp>\u003Ca href=\"https:\u002F\u002Fkyutai.org\u002Fpocket-tts\" rel=\"nofollow ugc noopener\">🔊 Demo\u003C\u002Fa> | \n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkyutai-labs\u002Fpocket-tts\" rel=\"nofollow ugc noopener\">🐱‍💻GitHub Repository\u003C\u002Fa> | \n\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Fpocket-tts\" rel=\"nofollow ugc noopener\">🤗 Hugging Face Model Card\u003C\u002Fa> | \n\u003Ca href=\"https:\u002F\u002Fkyutai.org\u002Fblog\u002F2026-01-13-pocket-tts\" rel=\"nofollow ugc noopener\">⚙️ Tech report\u003C\u002Fa> |\n\u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.06926\" rel=\"nofollow ugc noopener\">📄 Paper\u003C\u002Fa> | \n\u003Ca href=\"https:\u002F\u002Fkyutai-labs.github.io\u002Fpocket-tts\u002F\" rel=\"nofollow ugc noopener\">📚 Documentation\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch2>Main takeaways\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Runs on CPU\u003C\u002Fli>\n\u003Cli>Small model size, 100M parameters\u003C\u002Fli>\n\u003Cli>Audio streaming\u003C\u002Fli>\n\u003Cli>Low latency, ~200ms to get the first audio chunk\u003C\u002Fli>\n\u003Cli>Faster than real-time, ~6x real-time on a CPU of MacBook Air M4\u003C\u002Fli>\n\u003Cli>Uses only 2 CPU cores\u003C\u002Fli>\n\u003Cli>Python API and CLI\u003C\u002Fli>\n\u003Cli>Voice cloning\u003C\u002Fli>\n\u003Cli>Multi-language support: english, french, german, portuguese, italian, spanish\u003C\u002Fli>\n\u003Cli>Can handle infinitely long text inputs\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"#in-browser-implementations\" rel=\"nofollow ugc noopener\">Can run on client-side in the browser\u003C\u002Fa>\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Additional languages may be added in the future.\u003C\u002Fp>\n\u003Ch2>Trying it from the website, without installing anything\u003C\u002Fh2>\n\u003Cp>Navigate to the \u003Ca href=\"https:\u002F\u002Fkyutai.org\u002Fpocket-tts\" rel=\"nofollow ugc noopener\">Kyutai website\u003C\u002Fa> to try it out directly in your browser. You can input text, select different voices, and generate speech without any installation.\u003C\u002Fp>\n\u003Ch2>Trying it with the CLI\u003C\u002Fh2>\n\u003Ch3>The \u003Ccode>generate\u003C\u002Fcode> command\u003C\u002Fh3>\n\u003Cp>You can use pocket-tts directly from the command line. We recommend using\n\u003Ccode>uv\u003C\u002Fcode> as it installs any dependencies on the fly in an isolated environment (uv installation instructions \u003Ca href=\"https:\u002F\u002Fdocs.astral.sh\u002Fuv\u002Fgetting-started\u002Finstallation\u002F#standalone-installer\" rel=\"nofollow ugc noopener\">here\u003C\u002Fa>).\nYou can also use \u003Ccode>pip install pocket-tts\u003C\u002Fcode> to install it manually.\u003C\u002Fp>\n\u003Cp>This will generate a wav file \u003Ccode>.\u002Ftts_output.wav\u003C\u002Fcode> saying the default text with the default voice, and display some speed statistics.\u003C\u002Fp>\n\u003Cpre>\u003Ccode class=\"language-bash\">uvx pocket-tts generate\n# or if you installed it manually with pip:\npocket-tts generate\n\u003C\u002Fcode>\u003C\u002Fpre>\n\u003Cp>Modify the voice with \u003Ccode>--voice\u003C\u002Fcode> and the text with \u003Ccode>--text\u003C\u002Fcode>. We provide a small catalog of voices.\nChoose a pretrained language model with \u003Ccode>--language\u003C\u002Fcode> when running \u003Ccode>generate\u003C\u002Fcode>, \u003Ccode>export-voice\u003C\u002Fcode>, or \u003Ccode>serve\u003C\u002Fcode> (default: \u003Ccode>english\u003C\u002Fcode>). Non-english languages have also biggers 24 layers variants that are higher quality but slower. You can select them by using for example \u003Ccode>--language italian_24l\u003C\u002Fcode>.\nThe \u003Ccode>--config\u003C\u002Fcode> option accepts only a local YAML path for custom weights.\u003C\u002Fp>\n\u003Cp>You can take a look at \u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\" rel=\"nofollow ugc noopener\">this page\u003C\u002Fa> which details the licenses\nfor each voice.\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Falba-mackenna\u002Fcasual.wav\" rel=\"nofollow ugc noopener\">alba\u003C\u002Fa> (en)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Fpocket-tts\u002Fblob\u002Fadd_lang_not_documented\u002Fcommon_voice_it_36520747-enhanced-v2.mp3\" rel=\"nofollow ugc noopener\">giovanni\u003C\u002Fa> (it)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Fpocket-tts\u002Fblob\u002Fadd_lang_not_documented\u002Fcommon_voice_es_19762977-enhanced-v2.mp3\" rel=\"nofollow ugc noopener\">lola\u003C\u002Fa> (es)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Fpocket-tts\u002Fblob\u002Fadd_lang_not_documented\u002Fde-DE-juergen.mp3\" rel=\"nofollow ugc noopener\">juergen\u003C\u002Fa> (de)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Fpocket-tts\u002Fblob\u002Fadd_lang_not_documented\u002Fg-Vi8PgmSY0-enhanced-v2.wav\" rel=\"nofollow ugc noopener\">rafael\u003C\u002Fa> (pt)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Funmute-prod-website\u002Fdeveloppeuse-3.wav\" rel=\"nofollow ugc noopener\">estelle\u003C\u002Fa> (fr)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Fvctk\u002Fp228_023_enhanced.wav\" rel=\"nofollow ugc noopener\">anna\u003C\u002Fa> (en)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Fvctk\u002Fp303_023_enhanced.wav\" rel=\"nofollow ugc noopener\">azelma\u003C\u002Fa> (en)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Fvoice-zero\u002Fbill_boerst.wav\" rel=\"nofollow ugc noopener\">bill_boerst\u003C\u002Fa> (en)\u003C\u002Fli>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Fvoice-zero\u002Fcaro_davy.wav\" rel=\"nofollow ugc noopener\">caro_davy\u003C\u002Fa> (en)\u003C\u002Fli>\n\u003Cli>[charles](\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Fvctk\u002Fp254\" rel=\"nofollow ugc noopener\">https:\u002F\u002Fhuggingface.co\u002Fkyutai\u002Ftts-voices\u002Fblob\u002Fmain\u002Fvctk\u002Fp254\u003C\u002Fa>\u003C\u002Fli>\n\u003C\u002Ful>\n",1784240407896]