Clonar voz
A fully local, Spanish-language voice cloning and text-to-speech web application that turns a short recorded voice sample into speech reading any typed text, with no account, API or audio upload to a remote server.
Clonar voz records or accepts an uploaded ten-second voice sample, then generates speech in that same voice for any text a user types, running the entire process on the user's own computer with no server-side audio upload. It uses the Qwen3-TTS-12Hz-1.7B model in GGUF format, run through llama.cpp, and supports ten languages including Spanish, English, Chinese, German, Italian, Portuguese, Japanese, Korean, French and Russian.
It runs on either CPU or GPU, detecting and preferring a dedicated GPU when available, keeps a library of saved voices for reuse, splits long texts into sentences that are synthesized and rejoined automatically, shows live generation progress, and keeps a history of everything generated with playback and download. Once installed, it works with the network disabled.
Setup requires Python 3.10 or newer, a build of llama.cpp from August 2026 or later for Qwen3-TTS support, and ffmpeg for audio conversion; the model weights, roughly 1.5 GB, download automatically on first launch. It is written in Python.