Key Features
openEuler Intelligence is a comprehensive large language model (LLM) platform built on openEuler. It offers fundamental capabilities for integrating with OpenAI APIs and creating custom local knowledge bases. The platform also enables users to register semantic interfaces and Model Context Protocol (MCP) services, and develop Agent-based applications and workflows. Available in both web and client versions, it delivers exceptional convenience for developers and enterprises alike.
For developers:
A flexible, lightweight AI platform that enables the creation of personalized intelligent assistants to address everyday challenges
For enterprises:
A centralized intelligent agent solution that enhances traditional ISV and OSV software through workflow and Agent applications, significantly reducing costs associated with secondary development
RAG-based knowledge Q&A
The knowledge base module supports parsing mainstream document formats such as PDF, DOCX, and PPTX with straightforward operation.
Enterprises and individual developers can build personalized intelligent Q&A assistants in just one day by leveraging the knowledge base, local documents, built-in knowledge bases, and LLMs.
Workflow applications
Enterprises and individuals can upload local API definitions through OpenAPI, then visually orchestrate workflows to create applications that intelligently and precisely schedule API capabilities. This approach significantly reduces secondary development costs while facilitating intelligent upgrades for traditional software solutions.
Agent applications
Enterprises and individuals can register local MCP services and easily organize them to build Agent applications. These applications enable intelligent orchestration and scheduling of MCP services to accomplish complex tasks in production environments.
Architecture
openEuler Intelligence application layer
Functioning as the central hub for user interaction with LLMs, this layer empowers enterprises and developers to build AI applications through intuitive visual interfaces and low-code development. It streamlines deployment across intelligent operations, smart office solutions, and data analytics scenarios, boosting productivity significantly.
Full-stack compatibility
openEuler Intelligence delivers complete compatibility with leading AI software stacks, including drivers, SDKs, training/inference frameworks, and models, while integrating seamlessly across the broader AI ecosystem. This end-to-end solution connects hardware infrastructure to the application layer, eliminates interoperability barriers, and establishes a robust foundation for AI application implementation.
Inference acceleration
openEuler Intelligence harnesses advanced technologies—heterogeneous computing power awareness, memory convergence, CPU acceleration, and intelligent computing coordination—to unlock maximum hardware potential. It boosts inference performance while enhancing computational efficiency and reducing power consumption, enabling enterprises to optimize costs and elevate operational excellence.
Feature Highlights
Low-code development
openEuler Intelligence enables users to register traditional local services and visually orchestrate them into workflows, dramatically reducing secondary development overhead.
High availability
openEuler Intelligence integrates LLMs with traditional algorithms for intelligent workflow orchestration. It manages workflows based on user intent and constraints while using LLMs to select optimal combinations, ensuring high availability across all services.
Lightweight architecture
openEuler Intelligence operates efficiently on modest hardware, employing a 7B-parameter model for function calls and a 14B-parameter model for result synthesis while maintaining exceptional accuracy. Future releases will enhance the underlying sysHAX software to support these model sizes with improved token processing performance.
openEuler compatibility
openEuler Intelligence integrates natively within the openEuler image, with package selections precisely aligned to the openEuler repository. This native compatibility delivers optimized performance and speed on openEuler.
Contact us by following the Intelligence SIG or email:
intelligence@openeuler.org
