Open Web 3.0 platforms increasingly operate as \emph{service ecosystems} (e.g., DeFi, DAOs, and decentralized social applications) where \emph{admission control} and \emph{account provisioning} must be delivered as an always-on service under bursty demand. Service operators face a fundamental tension: enforcing Sybil resistance (one-person-one-account) while preserving user privacy, yet keeping on-chain verification cost and admission latency predictable at scale. Existing credential-based ZK admission approaches typically require per-request on-chain verification, making the provisioning cost grow with the number of concurrent joiners. We present \textbf{ZK-AMS}, a scalable admission and provisioning layer that bridges real-world \emph{Personhood Credentials} to anonymous on-chain service accounts. ZK-AMS combines (i) zero-knowledge credential validation, (ii) a \emph{permissionless} batch submitter model, and (iii) a decentralized, privacy-preserving folding pipeline that uses Nova-style recursive aggregation together with multi-key homomorphic encryption, enabling batch settlement with \emph{constant} on-chain verification per batch. We implement ZK-AMS end-to-end on an Ethereum testbed and evaluate admission throughput, end-to-end latency, and gas consumption. Results show stable verification cost across batch sizes and substantially improved admission efficiency over non-recursive baselines, providing a practical and cost-predictable admission service for large-scale Web 3.0 communities.

翻译：开放的 Web 3.0 平台日益作为**服务生态系统**（例如 DeFi、DAO 和去中心化社交应用）运行，其中**准入控制**和**账户供应**必须在突发需求下作为始终在线的服务提供。服务运营方面临一个根本性的矛盾：在保护用户隐私的同时实现女巫攻击抵抗（一人一账户），同时保持链上验证成本和准入延迟在大规模下可预测。现有的基于凭证的零知识准入方法通常需要每次请求进行链上验证，导致供应成本随并发加入者数量增长。我们提出了 **ZK-AMS**，一个可扩展的准入与供应层，它将现实世界的**人格凭证**与匿名的链上服务账户相连接。ZK-AMS 结合了（i）零知识凭证验证，（ii）一种**无需许可**的批量提交者模型，以及（iii）一个去中心化、保护隐私的折叠流水线，该流水线使用 Nova 风格的递归聚合与多密钥同态加密相结合，实现了每批次**恒定**链上验证成本的批量结算。我们在以太坊测试平台上端到端实现了 ZK-AMS，并评估了准入吞吐量、端到端延迟和 Gas 消耗。结果表明，验证成本在不同批次规模下保持稳定，并且准入效率较非递归基线有显著提升，为大规模 Web 3.0 社区提供了一种实用且成本可预测的准入服务。