卖什么
Fast inference for voice
商业模式拆解
买家：谁掌握预算？
未核实Engineering leads at companies running voice agents at scale.
触发点：什么让他们现在购买？
Latency and cost of GPT-class models for voice; site cites 80 ms TTFT vs 600 ms for GPT-4.1.
工作流：什么输入变成什么输出？
Hopper trains STT, TTS and speech LLMs on a customer production calls, serves them, and improves them against live traffic.
经济账：省了什么，怎么收费？
Usage-based inference; site cites $0.40 per 1M input tokens vs $2 for GPT-4.1.
渠道：怎么接触到客户？
Sales/engineer-led (Talk to an Engineer), console signup, technical blog.
融资历史与报道
| 融资日期 | 轮次 | 金额 | 投资方 | 来源 |
|---|---|---|---|---|
| — | YC 2026 秋季 | — | Y Combinator | ycombinator.com |
档案汇总自公开报道和公司官网。商业模式部分是 MaxBC 对这些材料的解读，不是已确认的事实；未知之处留空，不作猜测。 更新于 2026-10-12.