DEEPEXI TECH (01384) has announced that, effective October 15, 2026, its DeepWorks Enterprise Agent Platform (DeepWorks) will end its application-based trial mode and open self-registration to all users.
Users will no longer need an invitation code and can access the platform's functions by downloading the client and registering an account (trial entry: https://deepworks.cn/). The main arrangements for this opening are as follows:
Personal Edition: After registration, users will by default experience the DeepWorks Personal Edition, with a certain amount of free trial usage provided to newly registered users.
Enterprise Edition: Enterprise users can inquire about the DeepWorks Enterprise Edition solution to further build an AI employee platform tailored to their own business needs.
Building on its existing enterprise solution delivery business, the group is further providing standardized cloud services through DeepWorks. Users can self-register and use relevant agent functions, with a certain amount of free trial usage currently available. Subsequently, users can continue using the services by purchasing credits or through other actual payment methods, with related services billed by token usage.
This arrangement allows the group to further expand, beyond enterprise project delivery, a fee model related to customers' continued use of platform services. Deployment, system integration, and other services for enterprise clients will be separately agreed upon under relevant contracts.
DeepWorks is an enterprise agent platform built by the group for individual and enterprise users, namely an end-cloud collaborative, local-first platform for building, publishing, and operating AI employees for individual and enterprise users. DeepWorks can compile enterprise knowledge into two knowledge planes—a knowledge base Wiki and an enterprise business ontology—support vertical-domain model onboarding, and directly connect to enterprises' existing business systems. Together, these three elements enable AI employees to truly integrate into enterprises' core production processes.
DeepWorks integrates large models, skills, tool connections, and enterprise knowledge capabilities, supporting users in task planning, tool invocation, and result generation through natural language, and applying them to business scenarios such as data analysis, daily office work, manufacturing, operational decision-making, and data governance. Currently, DeepWorks' main capabilities include:
1. AI-Native Workspace: Based on the workspace, it integrates local files, task context, and multi-model capabilities, supporting task planning, tool invocation, and result generation through natural language. It covers various enterprise work scenarios such as business requirement analysis, business planning and execution, data analysis, and daily office work, achieving a complete closed loop from requirement initiation to result delivery.
2. Knowledge Center and Collaboration: The enterprise knowledge foundation parses multimodal data through the Deepexi enterprise large model and unifies the representation of cross-modal data time-series signals, compiling them into a traceable knowledge base Wiki and further extracting the enterprise business ontology. Relying on the enterprise knowledge foundation and enterprise business ontology, the model can understand, master, and output the internal logic of different enterprise business scenarios, obtain execution results for complex tasks, continuously accumulate knowledge, internalize enterprise knowledge as model capability, and become an expert in relevant scenario domains. By connecting enterprise knowledge and business scenarios, it supports AI result sharing and team collaboration, forming a continuous work loop from task execution and knowledge accumulation to capability reuse.
3. My AI Colleagues: A unified entry point for vertical-domain models and industry applications. It provides AI experts in fields such as operations analysis, merchandise planning, equipment diagnostics, remote operations and maintenance, data governance, and project management, supporting on-demand summoning of professional roles and multi-role division of labor and collaboration through expert teams. The group's DeepSense industrial manufacturing expert, DataSense business operations expert, FDESense data governance expert, and others can all be deployed and run on the DeepWorks platform.
4. Open Skill Ecosystem: Supports the creation, installation, management, and sharing of Skills, enabling the reuse of professional capabilities through Skill Hub, and connecting mainstream collaborative office tools and enterprise business systems through expansion methods such as MCP, plug-ins, and CLI.
DeepWorks, together with the group's Deepexi enterprise large model and the Deepology enterprise ontology dataset and skills, jointly constitute the group's product system for enterprise artificial intelligence applications. At present, the Deepology cloud shared knowledge base has achieved deep integration with DeepWorks, enabling enterprise agents to extend from general office and collaboration capabilities to deep business areas such as manufacturing, operational decision-making, and data governance. By accessing enterprise knowledge and related industry skills, DeepWorks can support the accumulation and reuse of enterprise knowledge and connect related collaborative office tools and enterprise business systems, providing users with relevant agent functions and cloud services.
The board believes that DeepWorks' opening of external registration and related payment arrangements will help the group expand its user base, further develop a fee model related to customers' continued use of platform services on top of its existing enterprise solution delivery business, and advance the commercialization of DeepWorks. The actual effectiveness of this opening and its impact on the group's operating performance will depend on various factors, including the market environment, user acceptance, payment conversion, actual usage, and the group's subsequent operational arrangements.
Comments