AI News AI资讯 6h ago Updated 2h ago 更新于 2小时前 45

Clipto Raises $15M to Build the Memory Layer for the AI Era Clipto融资1500万美元,打造AI时代的记忆层

Clipto raised $15 million at a $250 million post-money valuation from investors including HSG, GL Ventures, EnvisionX Capital, and Palm Drive Capital The startup launched native Windows and Android apps alongside a new Model Context Protocol (MCP) enabling AI agents to query private media libraries Clipto uses proprietary on-device AI to create a searchable memory layer from personal video and audio without uploading files to the cloud The platform offers semantic search, transcription, summarie Clipto完成1500万美元融资,投后估值达2.5亿美元,投资方包括HSG、GL Ventures、EnvisionX Capital、Palm Drive Capital等 推出原生Windows和Android应用,并引入Model Context Protocol (MCP)支持AI代理安全搜索用户私人媒体库 采用端侧AI技术将个人视频和音频转化为可搜索的记忆层,提供语义搜索、转录、摘要等功能,与Adobe Premiere等工具集成 所有数据处理均在本地完成,不上传云端,确保用户隐私和数据安全 已服务超过3000万用户,包括Google、Apple、Meta等科技公司的专业人士

65
Hot 热度
60
Quality 质量
65
Impact 影响力

Analysis 深度分析

TL;DR

  • Clipto raised $15 million at a $250 million post-money valuation from investors including HSG, GL Ventures, EnvisionX Capital, and Palm Drive Capital
  • The startup launched native Windows and Android apps alongside a new Model Context Protocol (MCP) enabling AI agents to query private media libraries
  • Clipto uses proprietary on-device AI to create a searchable memory layer from personal video and audio without uploading files to the cloud
  • The platform offers semantic search, transcription, summaries, and deep integrations with tools like Adobe Premiere, serving over 30 million users globally
  • Founder Henry Kang positions the product as an "AI memory layer" rather than just a reasoning layer, addressing the gap in understanding accumulated personal media

Why It Matters

Clipto addresses a critical bottleneck in the AI ecosystem: the inability of generative models to effectively access and reason over users' vast personal media archives. By keeping data on-device while enabling semantic search and AI agent integration, it bridges the privacy-performance tradeoff that has long constrained enterprise and consumer AI adoption. The MCP launch also signals growing momentum for standardized agent-to-data protocols in the AI infrastructure stack.

Technical Details

  • On-device AI architecture: Proprietary models run locally on user devices, performing video understanding, semantic search, transcription, and summarization without cloud upload, ensuring data privacy and reducing bandwidth costs
  • Model Context Protocol (MCP) integration: New MCP support allows AI agents to securely query and search users' private video and audio libraries, enabling agent-mediated access to personal media without exposing raw files
  • Cross-platform availability: Native applications for Web, Mac, iOS, Windows, and Android, with deep plugin integration into Adobe Premiere and other video editing workflows
  • Semantic search at scale: Natural language querying across terabytes of unstructured video and audio, with results mapped to exact timecodes for precise clip retrieval
  • Structured video notes: Automatic tagging and organization of media into searchable, structured formats that transform disconnected footage into an AI-readable knowledge base

Industry Insight

  • The rise of on-device AI memory layers represents a strategic shift from cloud-dependent AI services toward privacy-first, local-first architectures—companies that solve the "personal data access" problem will capture significant enterprise and prosumer value
  • MCP adoption by Clipto reinforces the growing standardization around agent-to-data protocols; AI tool builders should evaluate MCP compatibility to ensure their products can integrate with emerging memory and search infrastructure
  • The $250 million valuation for a privacy-focused video search startup signals strong investor conviction that personal media organization is a high-value, underserved category—expect continued funding and product innovation at the intersection of local AI, media workflows, and agent integrations

TL;DR

  • Clipto完成1500万美元融资,投后估值达2.5亿美元,投资方包括HSG、GL Ventures、EnvisionX Capital、Palm Drive Capital等
  • 推出原生Windows和Android应用,并引入Model Context Protocol (MCP)支持AI代理安全搜索用户私人媒体库
  • 采用端侧AI技术将个人视频和音频转化为可搜索的记忆层,提供语义搜索、转录、摘要等功能,与Adobe Premiere等工具集成
  • 所有数据处理均在本地完成,不上传云端,确保用户隐私和数据安全
  • 已服务超过3000万用户,包括Google、Apple、Meta等科技公司的专业人士

为什么值得看

Clipto展示了端侧AI在个人媒体管理领域的实际应用,为AI代理提供了新的数据访问方式。这种隐私优先的技术方案为AI记忆层的发展提供了重要参考,反映了AI从"生成"向"理解"演进的趋势。

技术解析

Clipto采用端侧AI架构,在本地设备上完成视频和音频的索引与处理,无需上传云端。通过语义搜索、转录和摘要功能,将非结构化媒体转化为可搜索的知识库。MCP协议的引入使AI代理能够安全地访问用户私人媒体库,实现人与AI共同使用个人媒体数据。

行业启示

端侧AI正在成为隐私保护与智能化结合的重要方向,企业应关注本地化处理能力的建设。AI代理需要可靠的记忆层来理解用户的历史数据和偏好,视频/音频等非结构化数据的智能处理将成为下一个竞争焦点。隐私优先的技术方案将成为企业级AI应用的重要差异化优势。

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Agent Agent Funding 融资 Product Launch 产品发布 Multimodal 多模态 Speech 语音