Skip to content
SPECIFICTION
Get listed

VoxCPM

voxcpm.com

Free plan WebmacOSLinuxAPISelf-hosted
Visit voxcpm.com
Spec score 8.8 #24 of 233 in Text-to-speech tools

What it is

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.

Spec sheet

Product
VoxCPM2 is a tokenizer-free text-to-speech system that generates expressive multilingual speech.Read 8 Oct 2026 · github.com
Platforms
web,macos,linux,api,self_hostedRead 20 Sep 2026 · voxcpm.com
Deployment
The repository documents Python and CLI use, a local web demo, production serving, and on-device inference.Read 8 Oct 2026 · github.com
Integrations
Documented serving options include vLLM-Omni with an OpenAI-compatible speech endpoint and Nano-vLLM for concurrent requests.Read 8 Oct 2026 · github.com
Support
The project links to Feishu, Discord, and its documentation wiki for community discussion and support.Read 8 Oct 2026 · github.com
Headquarters
Beijing, ChinaRead 28 Sep 2026 · voxcpm.com
Maker
The site identifies Modelbest Technology Co., Ltd. and offers custom solutions through [email protected].Read 8 Oct 2026 · voxcpm.com
API access
YesRead 20 Sep 2026 · voxcpm.com
API integration
Its vLLM-Omni serving option exposes an OpenAI-compatible /v1/audio/speech endpoint.Read 8 Oct 2026 · github.com
Audio output
The model outputs 48kHz audio and accepts 16kHz reference audio.Read 8 Oct 2026 · github.com
Commercial use
The project states its code and weights use the Apache-2.0 license and are free for commercial use.Read 8 Oct 2026 · github.com
Export formats
WAVRead 20 Sep 2026 · voxcpm.com
Fine-tuning
It supports full and LoRA fine-tuning, with adaptation possible from 5–10 minutes of audio.Read 8 Oct 2026 · github.com
Languages
VoxCPM2 supports 30 languages and lists nine Chinese dialects.Read 8 Oct 2026 · github.com
License and cost
The model weights and code use the Apache-2.0 license, and the repository says they are free for commercial use.Read 8 Oct 2026 · github.com
Limitations
Voice Design and controllable cloning results can vary between runs, and the project says unsupported languages may require testing or fine-tuning.Read 8 Oct 2026 · github.com
Platform support
The README documents Apple Silicon MPS support and C++ inference on CPU, Metal, CUDA, or Vulkan.Read 8 Oct 2026 · github.com
Requirements
The documented Python installation requires Python 3.10 or later but below 3.13, PyTorch 2.5.0 or later, and CUDA 12.0 or later.Read 8 Oct 2026 · github.com
Responsible use
The project forbids using VoxCPM for impersonation, fraud, or disinformation and recommends labeling AI-generated content.Read 8 Oct 2026 · github.com
Streaming
It provides a streaming generation API, and the repository reports real-time factor as low as about 0.3 on an NVIDIA RTX 4090.Read 8 Oct 2026 · github.com
Voice cloning
It supports controllable voice cloning from a short reference clip and continuation cloning with reference audio and its transcript.Read 8 Oct 2026 · github.com
Voice design
It can create a voice from a natural-language description without reference audio.Read 8 Oct 2026 · github.com

Plans and prices

Free & Open Source Free No paid tiers or usage limits stated Read 8 Oct 2026 · voxcpm.com

Every VoxCPM plan and price →

Strong points and limits

Strong points

  • Supports 30 languages and nine Chinese dialects
  • Natural-language voice design without reference audio
  • Streaming API and OpenAI-compatible endpoint
  • Apache-2.0 license permits commercial use
  • Full and LoRA fine-tuning

Limits

  • Voice Design results can vary between runs
  • Cloning results can vary between runs
  • Unsupported languages may require fine-tuning
  • Installation requires specified Python, PyTorch and CUDA versions

Who it suits

VoxCPM suits developers needing self-hosted multilingual speech, voice design, cloning or fine-tuning. It also fits commercial deployments that can meet its documented Python, PyTorch and CUDA requirements.

Verdict

VoxCPM offers an open, self-hostable speech model with multilingual output, APIs and fine-tuning. Plan for variable voice results and verify language support and system requirements before deployment.

Where Specifiction places it

Alternatives to VoxCPMAll alternatives →