Core Components¶
Explore the fundamental building blocks of the Agent-Graph platform.
Overview¶
Agent-Graph is built on several core components that work together to provide a comprehensive multi-agent platform:
- Agent: Autonomous units that execute tasks
- Graph (Workflow): Orchestration layer for multi-agent coordination
- Model: LLM integration and management
- Memory: Persistent storage for agent and user data
- MCP: Model Context Protocol integration
- Built-in Tools: Rich toolset for agent operations
- Prompt Center: Centralized prompt management
- Conversation: Share and export conversations
- Multi-User & Team: Collaboration features
Component Overview¶
Agents¶
Agents are the core execution units in Agent-Graph. They can use tools, maintain memory, and collaborate with other agents to accomplish complex tasks.
Graphs (Workflows)¶
Graphs provide visual workflow orchestration, allowing you to design and execute multi-agent coordination patterns.
Models¶
Agent-Graph supports various LLM models. Configure and manage your AI models through the model management system.
Memory¶
Agent-Graph provides both short-term and long-term memory systems for agents and users.
MCP Integration¶
Integrate Model Context Protocol servers to extend agent capabilities with external tools and data sources.
Built-in Tools¶
Agent-Graph comes with a rich set of built-in tools for file operations, agent creation, graph design, and more.
Learn more about Built-in Tools →
Prompt Center¶
Centrally manage and organize prompts for use across your agents and workflows.
Conversation Features¶
Share conversations with teammates and export them for analysis or documentation.
Learn more about Conversations →
Multi-User & Team¶
Manage teams, create invite codes, and collaborate with multiple users.
Select a component from the navigation to dive deeper into its features and usage.