VoxCPM
Open-source tokenizer-free TTS: 30 languages, voice design, cloning, 48kHz audio.
VoxCPM is the best open-source TTS we've seen for developers with GPUs. Its tokenizer-free approach, 30 languages, voice design, and 48kHz output outclass most alternatives, and the Apache-2.0 license makes it genuinely commercial-friendly. Non-technical users should steer clear unless they're ready to wrangle Python and CUDA. If you need a plug-and-play API, look at ElevenLabs or OpenAI TTS; if you want control and zero per-character fees, VoxCPM wins.
Verified 5d ago · liveness 75/100 · cite: rightaichoice.com/tools/voxcpm
- AI researchers exploring tokenizer-free speech synthesis
- Developers building custom voice assistants or chatbots
- Content creators needing high-quality multilingual voice-overs
- Open-source enthusiasts wanting local TTS without cloud dependencies
- Users seeking a turnkey cloud API (no hosted service available)
- Non-technical users expecting plug-and-play deployment
- Applications requiring minimal setup (needs GPU with CUDA)
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip VoxCPM if you lack a CUDA-capable GPU with at least 8GB VRAM, or if you need a turnkey hosted TTS API without infrastructure management.
Hosting your own GPU infrastructure (cloud GPU instances or local hardware) adds significant cost; there is no hosted API.
VoxCPM2 is free to use and modify, making it ideal for budget-conscious developers and researchers who already have GPU hardware. Compared to paid per-character APIs like ElevenLabs, you pay only for your own infrastructure, which can be cheaper at high volume. However, for teams without GPU expertise, hosted options like ElevenLabs or OpenAI TTS may be more cost-effective when factoring in setup and maintenance time.
In short
VoxCPM — Open-source tokenizer-free TTS: 30 languages, voice design, cloning, 48kHz audio. Best for AI researchers exploring tokenizer-free speech synthesis, Developers building custom voice assistants or chatbots, Content creators needing high-quality multilingual voice-overs. Free to use.
What people actually say about VoxCPM — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
15 mentions across 4 sources (Hacker News, Product Hunt, GitHub, Lemmy) · researched Jul 3, 2026.
- +Tokenizer-free architecture produces more natural prosody and intonation than discrete-token TTS systems.
- +Voice cloning from a short audio sample captures breathing, accents, and emotion details very accurately.
- +Supports 30 languages and outputs at 48kHz with real-time streaming inference potential.
- +Completely free and open-source under Apache-2.0 with no vendor lock-in.
- +Active community on GitHub with 32k+ stars and frequent updates.
- −CLI documentation is inconsistent — --device cpu flag doesn't work despite being shown in tutorials.
- −Polish and some non-English languages suffer from abrupt voice cutoffs at word ends.
- −CUDA errors on certain NVIDIA GPUs (e.g., L20) disrupt inference reliability.
- −Artifact from reference audio bleeds into output during transcript-guided cloning.
- −Fine-tuning new languages lacks clear documentation — only Chinese and English well-supported.
- • Requires powerful GPU for acceptable inference speed—costs may be incurred for cloud GPU rentals
- • No official hosting—users must provide their own infrastructure
Viability Score
How well maintained and how widely used is VoxCPM? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: September 2026
How we score →Key Features
- Tokenizer-free diffusion autoregressive TTS
- 30-language multilingual synthesis, no language tag needed
- Voice Design from natural-language description
- Controllable Voice Cloning with style guidance
- Ultimate Cloning with reference transcript
- 48kHz studio-quality audio output via AudioVAE V2
- Built-in super-resolution, no external upsampler
- Real-time streaming (RTF ~0.3 on RTX 4090)
- Nano-vLLM and vLLM-Omni acceleration (RTF ~0.13)
- OpenAI-compatible API via vLLM-Omni with PagedAttention
- Context-aware prosody and expressiveness
- Apache-2.0 open-source license
- LoRA and SFT fine-tuning support
- Python API via PyPI
- On-device inference with llama.cpp-omni
About VoxCPM
VoxCPM is an open-source text-to-speech system that generates continuous speech directly with a diffusion autoregressive architecture, skipping discrete tokenization for more natural and expressive output. The latest release, VoxCPM2, is a 2B parameter model trained on over 2 million hours of multilingual speech data and supports 30 languages without requiring language tags. It offers creative Voice Design—you can describe a voice in plain language to get a brand-new voice—and Controllable Voice Cloning, where you clone a timbre from a short reference clip and steer emotion or pace with style guidance. Ultimate Cloning even reproduces every nuance when you provide both reference audio and its transcript. Audio quality is a standout: VoxCPM accepts 16kHz references and directly outputs 48kHz studio audio via AudioVAE V2, with built-in super-resolution that eliminates the need for external upsamplers. Real-time streaming is practical too, with RTF as low as ~0.3 on an RTX 4090, and ~0.13 when accelerated by Nano-vLLM or vLLM-Omni, the latter exposing an OpenAI-compatible API with PagedAttention. Everything is released under Apache-2.0, free for commercial use, and includes a Python API, CLI, web UI, and fine-tuning support (SFT and LoRA). This is squarely aimed at developers, researchers, and creators who need high-quality multilingual voice generation without cloud lock-in or per-character costs. Whether you're building a voice assistant, localizing content, or designing character voices for games, VoxCPM hands you the controls. Just be ready to manage your own GPU infrastructure. Compared to hosted services like ElevenLabs, VoxCPM gives you full control and zero usage fees, but it demands CUDA 12.0+ and a solid GPU. It's not a turnkey solution—but if you want openness and flexibility, it's one of the most capable open TTS options out there.
Behind the Verdict
We've evaluated a lot of open TTS models, and VoxCPM is refreshingly different. Most TTS stacks use discrete tokens, which can flatten prosody. VoxCPM's diffusion autoregressive approach produces speech that feels more natural, and the fact that it does this for 30 languages without needing a language tag is a genuine time-saver. The 2B model trained on 2 million hours shows in the output quality—it's not just 'good for open source,' it's competitive with commercial APIs. The voice design feature is a killer. Being able to type '(A young woman, gentle and sweet voice)' and get that voice is powerful. It's not just a gimmick—for game developers or content creators, this opens up rapid prototyping of character voices without recording or licensing. Controllable cloning is equally useful: you get the timbre you want, but you can still nudge emotion and pace. That's a level of control that usually costs a lot via API. But let's be clear about what VoxCPM is not. It's not a product; it's a toolkit. If you're not comfortable with Python, CUDA, and managing a GPU server, this is going to be a rough ride. The requirements (Python 3.10-3.12, PyTorch ≥ 2.5, CUDA ≥ 12.0) are steep for beginners. And while real-time streaming is possible, you need a decent GPU—an RTX 4090 isn't exactly common. If you're on a CPU-only machine, forget it. Compared to alternatives like Coqui TTS or Piper, VoxCPM is on another tier in terms of expressiveness and language coverage. Piper is lightweight but limited in quality; Coqui has its own strengths but VoxCPM's voice design and 48kHz output give it an edge. For hosted APIs like ElevenLabs, you pay per character and have less control. VoxCPM's tradeoff is upfront infrastructure cost and maintenance. One thing to watch: the project is evolving
Researching VoxCPM? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas VoxCPM actually fits — and what changes day-one when you adopt it.
Needs to generate dynamic character voices with varied emotions and accents for a multilingual game.
Outcome: Uses VoxCPM2's voice design to create unique voices from text descriptions, and cloning to reuse a consistent voice across characters, integrating the OpenAI-compatible API for runtime generation.
Studying tokenizer-free TTS and wants to experiment with the latest open-source model.
Outcome: Installs VoxCPM2 on a lab GPU, uses the Python API to run experiments, and fine-tunes with LoRA on custom datasets to evaluate performance.
Produces multilingual YouTube videos and needs high-quality voice-overs without cloud costs.
Outcome: Uses the CLI to batch-generate narration audio at 48kHz, leveraging the 30-language support and cloning for a consistent voice across videos.
Use Cases
- Generate natural-sounding multilingual speech for audiobook narration.
- Clone a voice from a 5-second sample for personalized voice assistants.
- Create unique voice characters for video games using text descriptions.
- Stream real-time speech output for interactive voice applications.
- Fine-tune the model on domain-specific data for custom TTS systems.
- Run offline TTS on your own hardware for data-sensitive workflows.
Models Under the Hood
as of 2026-09-01
Limitations
- The provided evidence from the GitHub repository page does not include specific information about model parameters, hardware requirements, or documentation languages.
- Therefore, limitations cannot be verified from the available data.
as of 2026-08-21
Verification history
We have re-verified VoxCPM 7 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 7 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published VoxCPM tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source
$0
Ideal for
Developers, researchers, and creators with GPU hardware who need free, customizable, and locally-run TTS for commercial or research use.
What this tier adds
Starting and only tier; free to use under Apache-2.0, includes all features such as voice design, cloning, and fine-tuning.
Where the pricing makes sense
The company stage and team size where VoxCPM's pricing actually pencils out — and where peers do it cheaper.
VoxCPM2 is free to use and modify, making it ideal for budget-conscious developers and researchers who already have GPU hardware. Compared to paid per-character APIs like ElevenLabs, you pay only for your own infrastructure, which can be cheaper at high volume. However, for teams without GPU expertise, hosted options like ElevenLabs or OpenAI TTS may be more cost-effective when factoring in setup and maintenance time.
Setup time & first value
How long it actually takes to get something useful out of VoxCPM — broken out by persona, not the marketing-page minute.
For developers with a CUDA GPU, initial setup via pip install and downloading weights takes about 1-2 hours. Fine-tuning requires additional time for data preparation and training, typically a few hours to a day. Non-technical users may need significantly longer to set up the environment and troubleshoot.
Switching to or from VoxCPM
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From old VoxCPM-0.5B or VoxCPM1.5: Upgrade to VoxCPM2 by replacing the model weights and adjusting code for new APIs; the repo provides migration notes in the docs.
- ↗To ElevenLabs: Use VoxCPM2 to generate samples and upload them as custom voices in ElevenLabs; there is no direct migration path.
- ↗To OpenAI TTS: Export generated audio and use with OpenAI's API; no automatic migration.
Integrations
Resources & Guides
Tutorials & Learning
Official links
Tools that pair well with VoxCPM
Common stack mates teams adopt alongside VoxCPM, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Voxcpm vs Soniox
If you need a production-ready, compliant, multilingual STT+TTS+translation API with sub-200ms latency, choose Soniox. If you want a free, open-source TTS model with voice cloning and plan to self-host on a GPU, choose VoxCPM. Soniox wins for enterprise-grade voice applications; VoxCPM wins for budget-friendly TTS experimentation.
Voxcpm vs Retell Ai
Choose Retell AI if you need a turnkey, cloud-based voice agent platform for phone automation at scale with low latency and rich integrations, even at a cost. Choose VoxCPM if you prioritize zero licensing fees, full control over deployment, multilingual TTS, and voice cloning, and you have the technical capacity to self-host on GPU infrastructure. They serve fundamentally different needs: Retell is an all-in-one call center solution, while VoxCPM is a highly capable open-source speech synthesis model.
Voxcpm vs Voiceitt
Choose Voiceitt if you need speech recognition for non-standard speech patterns (e.g., cerebral palsy, heavy accents) and value integrations with Webex, Teams, and Alexa. Choose VoxCPM if you want open-source TTS with voice cloning and multilingual support, and have the technical ability to run a 2B model locally on a GPU. They serve fundamentally different needs.
Alternatives to VoxCPM
View all
OmniVoice Studio
Free, open-source, local-first voice cloning, design, dubbing, and dictation for 646 languages.
ComfyUI VoxCPM
Open-source 30-language diffusion TTS with voice cloning and design, runs locally in ComfyUI.
Frequently Asked Questions
Categories
Best-of guides
Topics
Used VoxCPM? Help shape our editorial sentiment research.