MCloudBridge Unveils Ai 365 Unified AI Platform for Public, Education, and Research Institutions
MCloudBridge Unveils Ai 365 Unified AI Platform for Public, Education, and Research Institutions
MCloudBridge 推出 Ai 365 Unified AI Platform,讓公共部門、教育及研究機構按資料敏感度、權限和政策,選用本地部署 AI 或外部 SaaS 服務。平台以 G1、G2、G3 分級路由資料,支援本地處理敏感資料,並可按政策處理後將 G2 資料交予公有 LLM。平台亦支援管理不同任務的 AI agents,並按查詢、資料特性和機構政策選擇處理環境。

MCloudBridge, an integrated AI business platform provider, said on Oct. 13 that it has unveiled its Ai 365 Unified AI Platform targeting public-sector, education, and research institutions.
The platform allows organizations to selectively configure and use local on-premise AI and external service AI (SaaS) according to their security policies, data sensitivity, user permissions, and business requirements.
According to MCloudBridge, sensitive data such as personal information, administrative records, and research materials that cannot be transferred outside an institution can be stored on local storage. Document analysis, knowledge searches, retrieval-augmented generation (RAG), and AI inference can then be performed in an isolated environment using local large language models (LLMs) such as Llama, Qwen, and DeepSeek.
For tasks where external AI use is permitted, the platform can connect to public cloud storage and external LLMs from providers such as OpenAI and Anthropic (Claude). AI agents for different tasks can also be managed within a single environment.
MCloudBridge said public-sector, education, and research institutions can use the Ai 365 Unified AI Platform to maintain data sovereignty while selecting AI tools suited to their security classifications and business objectives. The platform routes data for AI processing according to its sensitivity, using the G1, G2, and G3 classifications for information assets: G1 for critical/confidential assets, G2 for internal business assets, and G3 for general/public assets.
Critical and confidential data classified as G1 can be stored, searched, and processed for inference entirely within an institution using local storage and local LLMs. For internal business data classified as G2, the original documents can be analyzed in a local environment and key information extracted. Depending on institutional policies, the extracted information can then be tokenized or pseudonymized before being sent to public LLMs. For public data classified as G3, external public LLMs can be utilized directly for search, translation, summarization, and document drafting.
The platform also applies policy-based LLM routing to select an appropriate processing environment based on user queries, data characteristics, access permissions, and institutional policies. MCloudBridge said this approach reduces dependence on any single AI model and allows organizations to adopt newly released AI engines while retaining their existing platforms and work environments.
來源:The Elec · thelec.net