• 亿邦公众号
    • APP下载

Google Cloud推出开放知识格式 适配AI Agent跨系统调用

作者:亿邦AI 文章来源:亿邦动力

2026-06-15 14:18

作者:亿邦AI 文章来源:亿邦动力

2026-06-15 14:18

邦小白快读

EN
全文速览

本文核心内容是Google Cloud在2026年6月推出了全新开放知识格式OKF,核心作用是实现AI Agent知识跨系统可移植调用,解决现有AI应用的知识调用痛点,核心干货内容如下:

1. OKF的基础背景:该格式基于OpenAI早期核心成员、当前Anthropic成员安德烈·卡帕西提出的LLM wiki模式开发,跳出了传统检索增强生成架构的限制,能解决过往AI知识无法积累、跨文档处理效率低的问题。

2. OKF v0.1版本的设计:将知识标准化为带YAML前言的Markdown文件,仅设置必填type字段,其余字段均可自由配置,不同概念通过Markdown链接形成知识图谱,可被任意常用编辑器、搜索工具读取索引。

3. 核心优势:遵循最小化与可移植原则,生成方与使用方完全解耦,适配所有云服务商、数据库及AI Agent框架,相关规格和代码已经在GitHub公开。

Google Cloud推出的OKF开放知识格式,对品牌商布局AI应用、梳理内部知识资产有较高参考价值,核心干货总结如下:

1. 匹配品牌知识管理痛点:当前多数品牌都存在内部知识碎片化问题,知识分散在不同系统和员工个人经验中,布局AI客服、内部智能助手等AI Agent应用时,普遍面临知识调用难、适配成本高的问题,OKF针对性解决这类痛点。

2. OKF的适配优势:OKF设计灵活,不需要品牌推翻现有知识系统重构,仅需按照标准整理就能实现跨系统调用,人工和AI生成的知识都能适配,成本门槛较低。

3. 落地参考充足:目前OKF规格代码已经开源,Google Cloud已经上线了多个参考实现和样本包,还更新了Knowledge Catalog支持OKF导入,品牌可以直接测试试用,探索AI应用落地。

OKF的推出给AI相关领域的卖家指明了新的市场机会和发展方向,核心干货总结如下:

1. 明确存在的市场痛点:当前AI Agent开发赛道中,所有开发者都需要自行解决上下文适配问题,不同厂商重复开发同类数据模型,现有方案都是定制化开发互不兼容,知识锁定在自有系统内,市场存在大量未被满足的通用标准化需求。

2. 可切入的发展机会:卖家可以基于完全开放的OKF标准,开发适配不同行业场景的知识整理工具、AI Agent插件,面向存在知识碎片化问题的各类企业提供服务,切入空白市场。

3. 可利用的生态资源:OKF的规格和代码已经公开,Google Cloud已经给出了参考实现和Knowledge Catalog集成方案,卖家可以直接接入Google Cloud生态,降低开发和获客成本,同时需要注意跟进标准迭代,把握技术方向。

OKF的推出给工厂推进数字化和AI改造带来了新的思路和商业机会,核心干货总结如下:

1. 可参考解决工厂知识管理痛点:多数制造工厂同样存在知识碎片化问题,生产工艺、产品设计、设备维护等经验分散在老员工和不同系统中,推进AI质检、智能设计等AI Agent应用时,普遍面临知识调取难、适配成本高的问题,OKF的标准化思路可直接参考。

2. 改造成本低适配性强:OKF遵循最小化可移植原则,不需要工厂完全推翻现有数字化系统就能落地,不管人工整理的知识还是机器生成的知识都能适配,降低了中小工厂知识资产标准化的门槛。

3. 带来新商业机会:工厂可以按照OKF标准梳理自有知识资产,对接更多第三方AI应用,挖掘知识资产价值,同时对服务工厂数字化的服务商来说,也带来了标准化知识改造服务的新商机。

OKF的推出明确了AI服务领域的核心客户痛点,给出了通用化的解决方向,对各类To B AI服务商有较高参考价值,核心干货总结如下:

1. 清晰的客户痛点方向:当前所有企业布局AI Agent都面临知识碎片化问题,知识分散在多个存储位置,调用时需要拼接内容,不同厂商的定制方案互不兼容,知识被锁定在单一系统内,还存在大量重复开发的浪费,这类痛点都是服务商可切入的需求。

2. 可复用的标准方案:OKF是开放通用的知识标准,遵循最小化可移植设计,适配所有云服务商、数据库和Agent框架,所有规格和代码都已公开,服务商不需要从零开发数据模型,可直接基于OKF开发各行业的知识服务产品。

3. 可对接的成熟生态:Google Cloud已经上线了参考实现、样本包和成熟的Knowledge Catalog集成方案,服务商可以直接接入Google Cloud生态,降低开发和获客成本,快速推出落地产品。

OKF的推出反映了AI产业对开放知识标准的需求,给平台商的生态建设和运营指明了新方向,核心干货总结如下:

1. 明确平台生态的核心需求:当前AI Agent产业发展迅速,但知识碎片化、标准不统一、重复开发的问题严重限制了生态发展,开发者迫切需要通用开放的知识标准降低适配成本,平台围绕开放标准建设生态能吸引更多开发者入驻。

2. 可参考的平台落地做法:Google Cloud已经率先做出了实践,在推出OKF标准的同时,更新了自身的Knowledge Catalog功能,支持OKF导入并为AI Agent提供配套服务,还开放了所有规格和代码,这种开放生态的做法值得各类AI平台参考。

3. 需要跟进的趋势和需要规避的风险:如果平台继续坚持封闭定制化的知识方案,会逐步抬高开发者适配成本,导致开发者流失,平台可围绕OKF开放标准开展招商,吸引知识服务、AI Agent开发类商家入驻,丰富平台服务能力,跟上产业发展趋势。

OKF的推出是AI Agent知识管理领域的重要产业新动向,具备较高的研究价值,核心干货总结如下:

1. 产业新动向的背景:当前大模型和AI Agent产业快速发展,传统检索增强生成架构已经暴露出知识无法积累、跨文档综合处理效率低、知识无法跨系统移植的问题,行业迫切需要新的技术方案解决痛点,OKF是产业界应对该痛点的重要新尝试。

2. OKF的核心创新点:该标准跳出了传统架构的限制,基于安德烈·卡帕西提出的LLM wiki模式开发,采用最小化可移植的设计思路,实现了知识的跨系统互操作,解决了当前知识锁定、重复开发的行业痛点,是开放标准对抗封闭定制方案的新探索。

3. 可供研究的方向非常丰富:包括开放知识标准对AI Agent生态发展的影响、OKF的后续迭代方向、开放标准模式的商业化路径、对企业知识管理数字化转型的影响等,且Google公开了所有规格、代码和参考实现,为研究提供了充足的样本支撑。

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article covers Google Cloud's upcoming Open Knowledge Format (OKF), launching in June 2026. OKF is designed to enable portable, cross-system knowledge access for AI agents, addressing key pain points in knowledge invocation for current AI applications. Key takeaways are as follows:

1. Background: OKF is built on the LLM wiki concept first proposed by Andrej Karpathy, currently a core member at Anthropic and an early key contributor to OpenAI. It breaks free of the constraints of traditional Retrieval-Augmented Generation (RAG) architectures, and solves longstanding problems of non-cumulative knowledge and low cross-document processing efficiency.

2. OKF v0.1 design: OKF standardizes knowledge as Markdown files with YAML frontmatter, requiring only one mandatory "type" field while leaving all other fields fully configurable. Different concepts are connected via Markdown links to form a knowledge graph, and the format can be read and indexed by any common editor or search tool.

3. Core advantages: OKF adheres to principles of minimalism and portability, fully decoupling knowledge producers from knowledge consumers. It is compatible with all cloud providers, databases, and AI agent frameworks, and its full specifications and source code are already publicly available on GitHub.

Google Cloud's new Open Knowledge Format (OKF) offers valuable insights for brands building AI applications and organizing internal knowledge assets. Key takeaways for brands are as follows:

1. Aligns with core brand knowledge management pain points: Most brands today struggle with fragmented internal knowledge, scattered across disparate systems and locked in individual employee expertise. When building AI agent tools such as AI customer service or internal intelligent assistants, brands consistently face high integration costs and difficulties accessing knowledge across systems. OKF directly addresses these challenges.

2. Flexible integration advantages: OKF's flexible design does not require brands to completely rebuild their existing knowledge systems. Brands only need to organize existing knowledge to the open standard to enable cross-system access, supporting both human- and AI-generated content with a low cost barrier to adoption.

3. Readily available implementation resources: OKF's specifications and code are already open source, and Google Cloud has published multiple reference implementations and sample packages. Google Cloud has also updated its Knowledge Catalog to support OKF imports, allowing brands to test and explore AI application deployments directly.

The launch of OKF opens up new market opportunities and directions for sellers operating in AI-related fields. Key takeaways for sellers are as follows:

1. Clear unmet market demand: In the fast-growing AI agent development space, all developers currently have to build custom context integration solutions in-house. Different vendors repeatedly build identical data models, existing solutions are custom-built and incompatible with one another, and knowledge remains locked within proprietary systems. This leaves substantial unmet demand for a universal, standardized solution.

2. Attractive new market entry points: Sellers can build knowledge organization tools and AI agent plugins tailored to different industry use cases based on the fully open OKF standard. They can serve enterprises of all types struggling with knowledge fragmentation, addressing an entirely untapped market segment.

3. Leverage existing ecosystem resources: OKF's specifications and code are fully public, and Google Cloud provides pre-built reference implementations and a Knowledge Catalog integration solution. Sellers can integrate directly into the Google Cloud ecosystem to cut development and customer acquisition costs, while staying aligned with evolving standard specifications to keep pace with technical trends.

OKF introduces new approaches and business opportunities for manufacturers advancing digital and AI transformation. Key takeaways for factories are as follows:

1. A framework to solve common factory knowledge management pain points: Most manufacturing facilities also face fragmented knowledge: production processes, product designs, equipment maintenance expertise and more are scattered across veteran employees and disconnected systems. When deploying AI agent applications such as AI quality inspection or intelligent design, manufacturers consistently face high integration costs and difficulties accessing scattered knowledge. OKF's standardized approach offers a directly applicable solution.

2. Low transformation cost and broad compatibility: OKF follows minimalism and portability principles, and can be deployed without factories replacing their existing digital systems entirely. It supports both human-curated and machine-generated knowledge, lowering the barrier to knowledge asset standardization for small and medium-sized factories.

3. Unlocks new business opportunities: Factories can organize their proprietary knowledge assets to the OKF standard to integrate with more third-party AI applications, unlocking new value from existing knowledge. For digital service providers serving the manufacturing sector, OKF also creates new business opportunities for standardized knowledge transformation services.

OKF clarifies core customer pain points in the AI services space and outlines a universal solution direction, offering high value for all B2B AI service providers. Key takeaways for service providers are as follows:

1. Clearly defined customer pain points to address: All enterprises building AI agents today struggle with fragmented knowledge scattered across multiple storage locations, requiring custom content stitching for every invocation. Custom solutions from different vendors are incompatible, knowledge is locked within single systems, and massive resources are wasted on redundant development. All of these pain points represent addressable demand for service providers.

2. A reusable open standard: OKF is an open, universal knowledge standard built with minimalism and portability in mind, compatible with all cloud providers, databases, and AI agent frameworks. All of its specifications and code are publicly available, so providers do not need to build custom data models from scratch and can develop industry-specific knowledge service products directly on top of OKF.

3. Access to a mature, pre-built ecosystem: Google Cloud has already launched reference implementations, sample packages, and a production-grade Knowledge Catalog integration for OKF. Service providers can integrate directly into the Google Cloud ecosystem to reduce development and customer acquisition costs, and launch production-ready products quickly.

OKF reflects the AI industry's demand for open knowledge standards, and outlines a new direction for platform ecosystem building and operations. Key takeaways for platform operators are as follows:

1. Clear core demand for platform ecosystems: While the AI agent industry is growing rapidly, fragmented knowledge, inconsistent standards, and redundant development are severely constraining ecosystem growth. Developers urgently need a universal open knowledge standard to reduce integration costs, and platforms that build their ecosystems around open standards will attract more developer adoption.

2. A proven implementation model to follow: Google Cloud has already blazed a trail: alongside launching the OKF standard, it updated its Knowledge Catalog to support OKF imports and deliver supporting services for AI agents, and open-sourced all specifications and code. This open ecosystem approach serves as a valuable reference for all AI platforms.

3. Key trends to follow and risks to avoid: Platforms that stick to closed, custom knowledge solutions will gradually raise integration costs for developers and risk losing developer market share. Instead, platforms can onboard merchants focused on knowledge services and AI agent development around the OKF open standard, expand platform service capabilities, and align with broader industry trends.

OKF is a major new industry development in AI agent knowledge management with significant research value. Key takeaways for researchers are as follows:

1. Context for this new industry development: With the rapid growth of the large language model and AI agent industries, traditional RAG architectures have revealed clear flaws: they do not support cumulative knowledge, have low efficiency for cross-document processing, and cannot enable cross-system knowledge portability. The industry has an urgent need for new technical solutions to these pain points, and OKF represents a major new attempt by industry to address these challenges.

2. Core innovations of OKF: OKF breaks free of the constraints of traditional architectures. Built on Karpathy's LLM wiki concept, it adopts a minimal, portable design that enables cross-system knowledge interoperability and solves the industry-wide problems of knowledge lock-in and redundant development. It represents a new exploration of open standards as an alternative to closed custom solutions.

3. A wealth of promising research directions: These include the impact of open knowledge standards on AI agent ecosystem development, the future evolution trajectory of OKF, commercialization paths for open industry standards, and the impact of open standards on digital transformation of enterprise knowledge management. Google's public release of all OKF specifications, code, and reference implementations provides abundant empirical resources for research.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

2026年6月14日,Google Cloud推出全新开放知识格式,简称OKF。该规格将知识标准化为Markdown文件,实现跨系统可移植。该格式基于顶尖AI技术专家安德烈·卡帕西提出的LLM wiki模式开发为可互操作格式。卡帕西是OpenAI早期核心创始成员,曾主导特斯拉自动驾驶视觉技术研发,当前已加入AI巨头Anthropic。他提出的LLM wiki模式跳出传统检索增强生成架构限制,让大语言模型增量式构建并维护结构化、互联的Markdown文件集合,解决过往知识无法积累、跨文档综合处理效率低的问题。

OKF v0.1版本以带YAML前言的Markdown文件目录承载知识,仅设置必填的type字段,其余可选字段包含标题、描述、资源、标签、时间戳等,所有额外内容均可放入Markdown正文。不同概念通过标准Markdown链接关联形成知识图谱,OKF文件包可在任意编辑器读取,在GitHub完成渲染,也可被任意搜索工具索引。

OKF的推出指向当前企业普遍存在的知识碎片化问题。现有企业知识分散在元数据目录、wiki、代码注释、笔记本单元及工程师个人经验储备中,AI Agent调取特定信息时,需要从多个来源拼接内容片段。当前所有Agent开发者都需要自行解决上下文适配问题,不同目录供应商也在重复开发同类数据模型,现有同类解决方案均为定制化开发,互不兼容,知识始终被锁定在生成系统内部,这类痛点正是OKF的核心解决目标。

OKF整体设计遵循最小化与可移植原则,除必填的type字段外,文档类型、额外字段设置、正文结构均由生成方决定,生成方与使用方完全解耦。人工生成的OKF包可被AI Agent直接调用,机器生成的OKF包可在可视化工具中查看,适配所有云服务商、数据库及Agent框架。

伴随规格发布,Google Cloud同步上线多个参考实现,包含可爬取BigQuery数据集为每张表生成OKF文档的增强Agent、静态HTML可视化工具,以及GA4电商、Stack Overflow、比特币数据集三类样本包。Google Cloud同时更新Knowledge Catalog功能,支持导入OKF并向Agent提供相关服务,相关规格与代码已在GitHub公开,Knowledge Catalog集成方案单独公示。

【版权提示】亿邦动力倡导尊重与保护知识产权。未经许可,任何人不得复制、转载、或以其他方式使用本网站的内容。如发现本站文章存在版权问题,烦请提供版权疑问、身份证明、版权证明、联系方式等发邮件至run@ebrun.com,我们将及时沟通与处理。

已点赞

+1

{{ praise_text }}

好文章,值得被更多人看到

广告

下一篇

西班牙机器人公司Theker完成8500万美元A轮融资

西班牙AI机器人初创企业Theker完成8500万美元A轮融资,创欧洲纪录。其通用机器人可重构形态适配多场景,获三星等战略投资,旨在加速进入制造领域并扩大全球布局。

仓储物流 制造商 工业 制造业

2026-06-15 14:15:04