Gradium 发布 Voice Design：写一段描述文字，几秒生成全新合成语音

语音代理团队不断碰到同样的障碍。目录中有400个声音，而简报要求的是目录中没有的：蒙特利尔经销商的一位魁北克女接待员，一位六十多岁的叙述员，拥有讲堂般的权威感。简报的数量超过任何目录，而克隆则以每次一个说话者的方式弥合差距，每个都拥有来源、同意和许可。

Gradium：https://gradium.ai/，这家总部位于巴黎的语音AI公司由Kyutai研究实验室分拆出来，已经提供了另一种解决方案。Voice Design：https://gradium.ai/voice-design 可以读取书面描述，并在几秒钟内返回完整的新声音。无需参考音频，无需说话者，无需权利清除。

它可以部署吗？可以，Voice Design 已在 Gradium API 和 Studio：https://studio.gradium.ai/voices/design 上上线，包括免费套餐在内的每个计划都可使用，保留的声音在与目录声音相同的流式文本转语音端点运行，具有相同的延迟和输出格式。

描述是模型获得的唯一输入。Gradium 的文档：https://docs.gradium.ai/guides/voices/voice-design 列出了模型可以响应的属性，它们读起来像是选角通知：性别、年龄段、口音或出处、音高、语速、能量、音色和共鸣、声域和方式，以及声音正在执行的工作。描述长度为1到500个字符，可以使用英语、法语、西班牙语、葡萄牙语或德语。Gradium 建议以预期用途结尾，因为这会引导语音的表现和声域，而不仅仅是声音的颜色。

一次请求返回1到5个候选，通常在3到5秒内准备就绪。它们是单一角色的变体，因此不同的角色意味着不同的描述，而不是更多的样本。

流程包含四次调用。POST /voice-generator/generate 会生成候选 id 并设置 ready: false。GET /voice-generator/embeddings 会轮询直到状态改变。每个候选通过普通 TTS 端点进行试听，使用候选 id 作为 voice_id。POST /voices/from-embedding 会提升你选择保留的那个声音。

候选人有三项限制，而已转换的语音没有：试听文本限制为100个字符，他们仅能通过REST使用，且TTS WebSocket和语音对语音功能不支持他们。未转换的候选人在30天后会被删除。转换是免费的，可以清除过期时间，并使用一个与克隆共享的自定义语音槽。免费套餐可存放5个，自付套餐可存放1,000个。

采样是故意非确定性的。Gradium团队会首先扩展描述，这个扩展在每次请求中都不同，因此相同的提示即使使用固定种子仍会生成不同的语音。

Gradium对六种通过公共API可访问的语音设计系统和五种语言的口音提示进行了盲对比听力测试。母语者听两个未标记的音频片段，并选择更接近的匹配，或者判定为平局。在7,627次对比中，Gradium报告胜率为72.6%，比ElevenLabs的eleven_ttv_v3的59.0%高出13.6个百分点，随后是Inworld的44.8%、Fish Audio的36.7%和MiniMax的31.7%。胜率计算为胜利加半数平局，因此50%为基准。Gradium在五种语言中均位居第一。最显著的优势出现在多数目录处理过的区域口音：魁北克法语97%，里约拉普拉塔西班牙语86%，巴伐利亚德语85%，哥伦比亚西班牙语及非洲葡萄牙语83%。

基于相同提示集的模型评审结果一致。Gemini 3.1 Pro对每个未标记音频片段进行1到5的评分，并得出相同排序：Gradium 4.06，ElevenLabs 3.86，Inworld 3.64，Fish Audio 3.51。另据产品页面称，InstructTTSEval英文分集的提示遵循率为83.4%：https://arxiv.org/abs/2506.16381，这是TTS中遵循指令的学术基准。（注：所有这些数据均由供应商设计并运行。）

需要与我们合作推广您的GitHub仓库或Hugging Face页面或产品发布或网络研讨会等？请联系我们： https://forms.gle/wbash1wF6efRj8G58

Michal Sutter是一名数据科学专业人士，拥有帕多瓦大学数据科学硕士学位。凭借统计分析、机器学习和数据工程的扎实基础，Michal擅长将复杂的数据集转化为可操作的洞见。

以实践者为先的AI/ML新闻与分析，每月有超过100万开发者和研究人员阅读。

Voice agent teams keep hitting the same wall. The catalog holds 400 voices and the brief asks for the one that is not in it: a Quebecoise receptionist for a Montreal dealership, a narrator in his sixties with lecture hall authority. Briefs outnumber any catalog, and cloning closes the gap one speaker at a time, each carrying sourcing, consent and a licence.

Gradium：https://gradium.ai/, the Paris-based voice AI company spun out of the Kyutai research lab, has shipped a different answer. Voice Design：https://gradium.ai/voice-design reads a written description and returns complete new voices in a few seconds. No reference audio, no speaker, no rights to clear.

Is it deployable? Yes, Voice Design is live in the Gradium API and in Studio：https://studio.gradium.ai/voices/design, free on every plan including the free tier, and a kept voice runs on the same streaming Text-to-Speech endpoint as any catalog voice, at the same latency and output formats.

The description is the only input the model gets. Gradium’s documentation：https://docs.gradium.ai/guides/voices/voice-design lists the attributes it responds to, and they read like a casting call: gender, age band, accent or origin, pitch, pace, energy, timbre and resonance, register and manner, and the job the voice is doing. Descriptions run 1 to 500 characters in English, French, Spanish, Portuguese or German. Gradium advises ending with the intended use, because it steers delivery and register rather than only the colour of the voice.

One request returns 1 to 5 candidates, typically ready in 3 to 5 seconds. They are variations on a single character, so a different character means a different description, not more samples.

The flow is four calls. POST /voice-generator/generate mints candidate ids with ready: false . GET /voice-generator/embeddings polls until they flip. Each candidate auditions through the ordinary TTS endpoint, using the candidate id as voice_id . POST /voices/from-embedding promotes the one you keep.

Candidates carry three restrictions converted voices do not: audition text is capped at 100 characters, they are REST only, and the TTS WebSocket and Speech-to-Speech reject them. Unconverted candidates are deleted after 30 days. Converting is free, clears the expiry, and uses one custom voice slot shared with clones. The free tier holds 5, paid plans 1,000.

Sampling is deliberately non-deterministic. Gradium team expands the description first, and that expansion varies per request, so the same prompt with a fixed seed still yields a different voice.

Gradium ran a blind pairwise listening test on accent prompts across six voice design systems reachable through public APIs and five languages. Native speakers heard two unlabelled clips and picked the closer match, or a tie. Across 7,627 comparisons, Gradium reports a 72.6% win rate against the field, 13.6 points ahead of ElevenLabs eleven_ttv_v3 at 59.0%, followed by Inworld at 44.8%, Fish Audio at 36.7% and MiniMax at 31.7%. Win rate is wins plus half of ties, so 50% is par. Gradium placed first in all five languages. The widest margins came on regional accents that most catalogs flatten: Quebecois French at 97%, Rioplatense Spanish at 86%, Bavarian German at 85%, Colombian Spanish and African Portuguese at 83%.

A model judge over the same prompt set agreed. Gemini 3.1 Pro rated single unlabelled clips from 1 to 5 and produced the same ranking: Gradium 4.06, ElevenLabs 3.86, Inworld 3.64, Fish Audio 3.51. Separately, the product page claims 83.4% prompt adherence on the English split of InstructTTSEval：https://arxiv.org/abs/2506.16381, the academic benchmark for instruction following in TTS. (Note: All of these numbers are vendor designed and vendor run.)

Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us ：https://forms.gle/wbash1wF6efRj8G58

Michal Sutter is a data science professional with a Master of Science in Data Science from the University of Padova. With a solid foundation in statistical analysis, machine learning, and data engineering, Michal excels at transforming complex datasets into actionable insights.

Practitioner-first AI/ML news and analysis, read by 1M+ developers and researchers every month.