# KenduriLuhh: A Multi-Agent AI System Revolutionizing Traditional Feast Management in Malaysia

> KenduriLuhh is a multi-agent AI system built on Azure OpenAI and AutoGen, designed for intelligent catering management of traditional Malaysian feasts (kenduri/rewang), enabling end-to-end automation from menu planning to logistics coordination.

- 板块: [Openclaw Llm](https://www.zingnex.cn/en/forum/board/openclaw-llm)
- 发布时间: 2026-05-01T13:14:10.000Z
- 最近活动: 2026-05-01T13:23:28.827Z
- 热度: 159.8
- 关键词: 多智能体系统, AutoGen, Azure OpenAI, 餐饮管理, 马来西亚文化, 宴席规划, rewang, 本地化 AI
- 页面链接: https://www.zingnex.cn/en/forum/thread/kenduriluhh-ai
- Canonical: https://www.zingnex.cn/forum/thread/kenduriluhh-ai
- Markdown 来源: floors_fallback

---

## Introduction: KenduriLuhh—A Multi-Agent AI System Revolutionizing Traditional Feast Management in Malaysia

KenduriLuhh is a multi-agent AI system built on Azure OpenAI and AutoGen, specifically designed for traditional Malaysian feasts (kenduri/rewang). It addresses the pain points of traditional experience-driven manual management, enabling end-to-end automation from menu planning to logistics coordination. The system is deeply adapted to local culture, helping to intelligentize traditional feast management while inheriting and preserving the rewang culture of community mutual assistance in Malaysia.

## Challenges in Traditional Feast Management in Malaysia and the Cultural Background of Rewang

In Malaysia, kenduri (feasts) and rewang (collective mutual assistance activities) are at the core of community life, involving catering needs for hundreds of people. Traditional management relies on manual experience and faces issues such as unfamiliarity among the younger generation and complex coordination. Rewang originates from the mutual assistance spirit (gotong-royong) of the Malay community, where neighbors divide labor and cooperate to prepare food, but there are coordination difficulties such as portion synchronization, division of labor adjustments, and cost allocation. KenduriLuhh targets these pain points and provides an AI solution deeply adapted to local culture.

## Multi-Agent System Architecture of KenduriLuhh

The system is built based on the AutoGen framework and adopts a multi-agent collaboration model:
- **Menu Planning Agent**: Generates optimized menus based on event type, number of people, and budget, considering taste, seasonality, and procurement convenience;
- **Ingredient Procurement Agent**: Converts menus into procurement lists, marking prices, suppliers, and priorities;
- **Logistics Coordination Agent**: Manages volunteer division of labor, assigning tasks based on skills, time, and geographical location;
- **Budget Management Agent**: Tracks costs and proposes adjustment suggestions when over budget;
- **Quality Control Agent**: Monitors progress and quality, setting checkpoints to ensure food safety.

## Deep Localization Adaptation: Aligning with Malaysian Culture and Scenarios

The uniqueness of KenduriLuhh lies in its adaptation to local scenarios:
- **Cuisine Knowledge Base**: Built-in local dish data, including traditional recipes, regional variations, and batch production adaptations;
- **Cultural Sensitivity**: Automatically adapts to religious/cultural dietary taboos (halal, vegetarian, etc.);
- **Local Supply Chain**: Integrates Malaysian supplier and market data, providing location-related procurement suggestions;
- **Multilingual and Units**: Supports Malay, English, and Chinese, compatible with traditional measurements (kati, tahil) and metric units.

## Practical Workflow Demonstration: Case of a 200-Person Wedding Feast

Take a 200-person Malay wedding in Penang (budget of 5000 MYR) as an example:
1. **Requirement Collection**: Users input event information;
2. **Menu Generation**: Recommends traditional + innovative dishes such as rendang chicken and curry fish;
3. **Procurement List**: Generates a detailed list, suggesting bulk purchase of spices at Penang's Little India;
4. **Manpower Coordination**: Assigns tasks based on volunteer information (e.g., 5 people responsible for staple foods, 3 for preparing rendang);
5. **Budget Monitoring**: Shows an estimated expenditure of 4800 MYR, with emergency reserves set aside;
6. **Execution Monitoring**: Sends reminders and adjusts delay plans (e.g., simplifying decorations).

## Technical Implementation: Combined Application of Azure OpenAI + AutoGen

Technology selection:
- **Azure OpenAI GPT-4o**: Provides natural language understanding/generation capabilities, supporting multimodality (e.g., ingredient image recognition);
- **AutoGen Framework**: Open-source multi-agent dialogue framework, supporting group chat collaboration to simulate rewang scenarios;
- **Knowledge Graph**: Stores localized dish, ingredient, and supplier data;
- **Mobile Interface**: Responsive web and WhatsApp integration for convenient on-site use.

## Social Value and Future Development Directions

Social value: Preserves traditional feast culture and helps the younger generation participate in rewang; open-source community model supports joint improvement. Future plans:
- Carbon Footprint Calculation: Provides sustainable alternative solutions;
- Skill Inheritance Module: Records the cooking experience of the older generation;
- Cross-Community Collaboration: Supports resource sharing and mutual assistance networks.

## Conclusion: Technology Serving Tradition, Integration of Humanity and AI

KenduriLuhh demonstrates how AI can take root in specific cultural scenarios and solve coordination problems. It does not replace interpersonal interaction but reduces organizational burdens, allowing people to focus on the joy of gathering. Integration of technology and humanity: Advanced AI serves traditional community practices, but respects and protects cultural values, reflecting the humble service role that technology should have.