VoxKiln is a local, offline voice synthesis app for Windows. It turns text into natural speech using bundled generative AI models that run on your own device. Create new voices from style prompts, synthesize multilingual narration, or use authorized reference audio to guide voice style. By default, your text, reference audio, and generated audio are processed locally and are not uploaded to AILocalabs servers. VoxKiln is built for creators, developers, educators, and teams who need private, repeatable voice generation without depending on a cloud voice API. An NVIDIA CUDA-capable GPU is required for this offline Windows package. Initial startup may take several minutes while local models load. Learn more: https://ailocalabs.com/en/products/voxkiln/