|
V4.0
|
July 2025
|
Comprehensive AI Capability Upgrade
|
-
Launched the AI Knowledge Management module, supporting intelligent retrieval, automatic summarization, and knowledge agent applications.
-
Released the One-Stop Data Portal, enabling multi-source data integration and unified management.
-
Added Agent Store and Agent orchestration features to help users quickly build business agents.
-
Comprehensively strengthened foundational AI capabilities, improving the system's reasoning and collaboration performance.
|
|
V4.1
|
October 2025
|
AI Studio Enhancement and System Security Optimization
|
-
AI Studio (Agent module) was comprehensively upgraded, supporting personalized agent development and management.
-
Added token and permission management features to improve the system security framework and access control mechanisms.
-
Optimized AI translation and glossary management, improving industry-level translation accuracy.
-
Agent orchestration capabilities were enhanced, supporting multi-agent collaboration for executing complex tasks.
|
|
V4.2
|
July 2026
|
Focused on Enterprise-Grade AI Application Scale-Up, with Comprehensive Upgrades Across Five Directions: Agent Execution, Knowledge Services, Data Ingestion, Workflow Orchestration, and Platform Governance
|
Super Agent Execution Capability
-
The core feature of AI Central 4.2, designed to aggregate multiple dispersed Agent capabilities within the platform into a unified conversation entry point, forming a continuously collaborative task execution flow. Its key advantage is autonomously decomposing tasks and invoking Agents, Skills, and Tools to collaboratively complete work.
New Out-of-the-Box Tool Experiences
-
AI Slides: AI-powered automatic PPT generation
-
AI Translator: Supports online translation of documents (PPT, PDF, scanned PDF, Word, Excel) and paragraph-level text translation
-
Deep Research: Enterprise-grade "research and analysis engine" that automatically generates in-depth research reports
Enterprise Knowledge Base and Pipeline Knowledge Engineering
-
Launched the preprocessing Pipeline, supporting customized parsing, chunking, and vectorization workflows by file type.
-
Added MCP protocol integration, enabling standardized access to the Model Context Protocol.
-
Strengthened knowledge base retrieval capabilities, optimizing chunking strategies and field extraction accuracy.
-
Improved the plugin and skill extension system, supporting custom plugin development and management.
Data Ingestion and Analytics Capability
Bug Fixes and Other Known Issues Resolved
|
