字节前高管创业Apex Context获千万融资

作者:cooldora 发布于:2025年10月31日 阅读量:4
字节跳动前AI产品副总裁李谦离职创立"Apex Context",定位为营销场景多模态Agent。公司成立仅两周即获得硅谷与百度风投数百美元种子轮融资。其核心自研长窗口多模态模型可一次性处理10万字品牌资料,自动生成4K视频及可编辑分层文件,支持Adobe/Figma直接编辑。该产品通过理解隐性需求替代表单填写,内部测试已为国际饮料巨头及本土美妆集团节省70%外包成本,项目周期从3周缩短至48小时。公司采用效果计费模式:仅当CPM低于行业均值或达成约定点击率时收取

Li Qian, former vice president of AI products at ByteDance, has left the company and founded "Apex Context," positioning it as a multimodal Agent for marketing scenarios. Just two weeks after its establishment, the company secured a seed round of several million dollars from Silicon Valley's HT Investment and Baidu Ventures, with the valuation not disclosed.

Li Qian once led multiple AI product lines at Shengshu Technology. He compared his new company to the "007 logistics department" — when a brand provides a brief, the Agent automatically completes strategy, scripts, storyboards, video generation, and delivery material packages, requiring only a nod or shake from the client. The core technology is a self-developed long-window multimodal model, which can process up to 100,000 words of brand manuals and historical materials at once, generating 4K videos and editable layer files that can be directly opened and further processed in Adobe and Figma.

"Less control, more context," Li Qian explained the origin of the company's name: allowing the model to understand implicit needs like an experienced agency, rather than having users fill out forms. Internal test clients already include an international beverage giant and two local beauty groups, saving an average of 70% on outsourcing costs and reducing project delivery cycles from 3 weeks to 48 hours.

Apex Context uses a results-based billing model: it starts taking commissions only if CPM is lower than the industry average. If the campaign does not meet the agreed CTR, the platform will forgo sharing the profits. Li Qian plans to open 100 industry Agent templates within 3 months, with education, short dramas, and live streaming scripts also launching. "Big companies are busy competing in general large models, which is the opportunity window for startups to seize scenarios."

相关文章

暂无相关文章

评论区 (加载中...)

加载评论中...