Voicebox vs VoxCPM

Voicebox is a free, open-source local AI voice studio for voice cloning, speech generation, dictation, and agent voice output. Its desktop app can clone a voice from as little as three seconds of audio, supplied by file, microphone recording, or system capture. The maker lists seven text-to-speech engines for cloning, preset voices, and delivery control. A timeline editor supports multi-voice narratives, arranging tracks, trimming clips, and mixing conversations; effects include pitch shift, reverb, delay, and compression. Dictation uses a global shortcut to send transcripts to the focused text field or clipboard, with local Whisper models supporting 99 languages. MCP-aware agents such as Claude Code, Cursor, and Cline can speak through Voicebox, and a local REST API provides speech generation and related functions. The privacy policy says the offline desktop app requires no account and does not send recordings, profiles, transcripts, generations, or settings to the maker. Local is free with unlimited local generations and captures. Cloud and mobile sync are marked as coming soon, and cloud pricing and limits are not final.

VoxCPM is an open-source text-to-speech system whose VoxCPM2 model generates expressive multilingual speech. It supports 30 languages and lists nine Chinese dialects. Voice Design can create a voice from a natural-language description without reference audio, and voice cloning is supported. Output is 48kHz audio, reference audio is accepted at 16kHz, and WAV is listed as an export format. A streaming generation API is available; the repository reports a real-time factor as low as about 0.3 on an NVIDIA RTX 4090. Full and LoRA fine-tuning can adapt the model from 5–10 minutes of audio. Documented deployment includes Python and CLI use, a local web demo, production serving and on-device inference. vLLM-Omni exposes an OpenAI-compatible /v1/audio/speech endpoint, while Nano-vLLM supports concurrent requests. Apple Silicon MPS and C++ inference on CPU, Metal, CUDA or Vulkan are documented. Code and weights use Apache-2.0 and are free for commercial use. Python installation requires Python 3.10 or later but below 3.13, PyTorch 2.5.0 or later and CUDA 12.0 or later. Voice Design and cloning results can vary, and unsupported languages may require testing or fine-tuning. VoxCPM is headquartered in Beijing, China.