天枢:管理与交付模型TianShu: manage and deliver models
模型资产管理与服务平台Model asset management and services
管理模型资产和目录,包装、交付与部署模型服务或模型包,并观察其运行情况。Manage model assets and catalogs, package, deliver and deploy model services or packages, and observe their operation.
协调云端任务、连接真实现场、验证设备能力,分别由引核、引元与引芯承担。模型供给、经验积累和能力发现,为这三类运行职责提供支持。Cortex coordinates cloud tasks, Origin connects physical spaces and Nexus establishes device trust. Model supply, experience and capability discovery support these runtime responsibilities.
产品职责与协作方式正在具体项目中验证;可用能力和接入范围按项目确认。Product responsibilities and coordination are being validated through specific projects. Available capabilities and integrations are confirmed per project.
职责与机制示意Roles and mechanics
任务、权限与能力调用Tasks, permissions and capability calls
承载边缘运行,连接现场Hosts edge operation and connects the field
在允许的边界内执行Act within permitted boundaries
身份 · 位置 · 能力Identity · location · capabilities
实际状态与结果,回到任务核查Actual state and results return for verification
需要安排云端任务、跟进全局状态或协调多个空间?引核组织云侧智能体运行、权限策略、模型与能力调用,以及失败处理。Need to organize cloud tasks, track global state or coordinate several spaces? Cortex handles cloud agent runtime, permission policies, model and capability calls, and failures.
了解引核Explore Cortex需要把设备接入真实空间,并确认动作后的变化?引元连接现场,维护空间状态,组织边缘运行与反馈。Need to connect devices in a physical space and check what changes after action? Origin connects the field, maintains spatial state and supports edge operation and feedback.
了解引元Explore Origin需要确认设备是谁、在哪里、能做什么?引芯验证设备身份与能力,明确可信连接和允许调用的边界。Need to establish who a device is, where it is and what it can do? Nexus verifies device identity and capabilities, with trusted connections and clear calling boundaries.
了解引芯Explore Nexus天枢管理模型服务与模型包;各模型维护自身质量与版本。模型给出的判断,需要经过现场状态、权限与规则校验。TianShu manages model services and packages. Each model maintains its own quality and versions. Model judgments need checks against field state, permissions and rules.
模型资产管理与服务平台Model asset management and services
管理模型资产和目录,包装、交付与部署模型服务或模型包,并观察其运行情况。Manage model assets and catalogs, package, deliver and deploy model services or packages, and observe their operation.
模型能力与生命周期Model capabilities and lifecycle
灵枢面向认知,地枢面向物理空间感知与状态理解。各模型分别负责自身训练、评测、质量和版本。LingShu focuses on cognition; DiShu on physical-space perception and state understanding. Each model owns its training, evaluation, quality and versions.
元域负责能力成长,织界连接需求与贡献,开发者平台提供能力供给入口。完整的能力交易、认证与结算网络仍是后续建设方向。Realm supports capability growth, Orcha connects demand and contributions, and the developer platform provides a supply entry. A complete network for capability trading, certification and settlement remains a future direction.
长期经验与能力版本Long-term experience and capability versions
保存任务经验与来源记录,评估和维护能力版本,让经过验证的改进有机会进入下一次任务。Keep task experience and provenance, evaluate and maintain capability versions, and bring validated improvements into later tasks.
探索元域Explore Realm能力发现与贡献记录Capability discovery and contribution records
组织能力发现、供需匹配、调用与贡献记录;贡献计量与价值分配的具体机制,随运行证据和建设阶段确认。Support discovery, matching, calls and contribution records. Operating evidence and development stages inform the mechanisms for contribution measurement and value distribution.
探索织界Explore Orcha面向开发者的供给入口Supply entry for developers
了解能力封装、测试与发布路径,以及如何连接元域和织界。可用接口与接入要求以平台当前文档为准。Explore packaging, testing and publishing paths, and connections to Realm and Orcha. Current platform documentation determines available interfaces and integration requirements.
进入开发者平台Open developer platform有业务场景,先明确需要改善的任务和验证标准;有设备、软件服务或智能体能力,再查看相应接入要求。For a business scenario, define the task to improve and validation criteria. For a device, service or agent capability, explore the relevant integration requirements.