What is Gendial
Gendial is the homophone of "煎蛋" (Gendial), which stands for General Dialog Agent, a general-purpose conversational AI platform.
Gendial is a programming platform for creating AI-native intelligent agents. Here, you can create various AI-native agents, ranging from no-code to high-code, from intelligent customer service to full-process automation tools.
No-Code: On the Gendial platform, you can create your own AI agent by simply dragging and dropping, without writing a single line of code. It's as simple as making an egg.
High-Code: You can also use the code nodes, API call nodes, and various logic judgment nodes provided by the platform to implement complex business logic and create AI-native applications.
Advantages of Gendial Platform
The Gendial platform is mainly aimed at enterprise users, dedicated to creating intelligent agents that can handle real business scenarios. Compared to other intelligent agent creation platforms, Gendial has the following advantages:
1. No-Code / High-Code Intelligent Agent Creation
The flexible and convenient way of creating intelligent agents allows businesses to have complete control over their end-to-end business processes. The platform comes with various pre-built system nodes that can satisfy most logic control needs, greatly reducing the amount of prompt engineering required. It also supports full customization of prompts (via blank nodes) and the ability to implement code logic, allowing for the creation of complex business logic.
2. Native Multi-Agent Support
On the Gendial platform, you can break down a complex enterprise task into multiple intelligent agents that collaborate with each other. Multiple agents can take on different roles, work together, and supervise each other. Additionally, enterprise standard procedures can be made into a single agent, which can be called by multiple agents, avoiding redundant effort in recreating the same functionality.
3. Powerful Knowledge Base System
The Gendial platform includes a powerful and user-friendly knowledge base system, supporting the configuration of complex RAG (Retrieval-Augmented Generation) strategies to meet the needs of different document types and formats. It also supports a single intelligent agent connecting to multiple knowledge bases and configuring AI selection strategies, decoupling the agent's logic from the knowledge base (data), making it easier to reuse the agent workflow. The knowledge base supports both standard vector databases based on embedding models and file libraries that use large language models for retrieval, allowing flexible choices for various scenarios.
4. Integration with Enterprise Business Systems
Using the API call nodes and code nodes on the Gendial platform, intelligent agents can easily integrate with existing enterprise business systems, including but not limited to enterprise databases, OA systems, email systems, etc. Agents can retrieve real-time business data from these systems to better answer user queries or automatically initiate or flow business processes based on user needs, truly integrating AI capabilities with enterprise business functions.
5. Enterprise-Level Permission Control
The Gendial platform includes a comprehensive permission management system, allowing the creation of user groups, role assignments, and the implementation of permission isolation between different business departments. Both individual agents and knowledge bases can have fine-grained permission control (read, write, use, etc.), meeting enterprise security management needs. The platform also supports integration with existing enterprise unified identity authentication systems, such as CAS, oAuth2, or Microsoft Azure AAD login, allowing enterprise users to log in seamlessly.
6. Multi-Channel Access
The intelligent agents created by users can be easily published to various platforms via H5 client, such as WeChat, WeCom, Feishu, Teams, and more. The platform also supports OpenAI-compatible API calls, allowing perfect integration with any client or application that supports OpenAI. The platform has a built-in Intelligent Agent Marketplace where users within the same tenant can easily find various useful intelligent agents. Published agents can also specify authorization scopes and provide fine-grained control over user permissions.