安全合规Security & Compliance

安全理念Security Philosophy

我们将安全视为产品的底层属性而非附加功能。从架构设计、代码开发到运营运维，全流程遵循"安全左移"与"最小权限"原则，默认启用加密与审计能力。Security is a foundational property of our products, not a feature. We follow shift-left security and least-privilege principles across design, development and operations.

数据安全Data Security

我们提供全私有化部署选项，客户数据不出域。传输层采用 TLS 1.3 加密，存储层支持透明加密与密钥自管理（BYOK）。敏感信息默认脱敏展示，操作留痕全量审计。Fully on-prem deployment keeps data in your domain. Transits use TLS 1.3; storage supports encryption and customer-managed keys (BYOK). Sensitive data is masked by default with full audit trails.

模型安全Model Security

针对大模型场景，我们构建了覆盖输入输出双向的内容安全护栏：输入侧抵御提示注入与越狱攻击，输出侧进行违法与侵权内容过滤，全程留存审计日志，支持按监管要求回溯。For LLM workloads we provide bidirectional guardrails: prompt-injection and jailbreak defenses on input, illegal-content filtering on output, with auditable logs for regulatory traceability.

合规体系Compliance

我们的产品与服务按以下标准体系设计与交付：undefined

• 网络安全等级保护（三级）参考设计，支持密评；MLPS Level 3 aligned design with crypto evaluation support.
• 信创环境全栈适配（芯片、操作系统、数据库）；Full Xinchuang stack adaptation.
• 生成式人工智能服务相关备案与安全评估流程支持；Support for generative AI filing and safety assessment processes.
• 数据出境合规咨询与方案支持。Data export compliance consulting and solution support.

漏洞报告Vulnerability Disclosure

如您发现我们产品或服务的安全漏洞，请发送邮件至 security@taioneai.com。我们承诺在 3 个工作日内响应，并对有效报告者致谢。请勿利用漏洞进行超出验证必要范围的测试。Report vulnerabilities to security@taioneai.com. We respond within 3 business days and credit valid reports. Please avoid testing beyond what is necessary for verification.