graph LR
CLI_Deployment["CLI & Deployment"]
Orchestration_Core["Orchestration Core"]
Agent_System["Agent System"]
Task_Execution_Layer["Task Execution Layer"]
LLM_Tool_Integration["LLM & Tool Integration"]
Memory_Knowledge_Base["Memory & Knowledge Base"]
Observability_Telemetry["Observability & Telemetry"]
CLI_Deployment -- "initiates execution within" --> Orchestration_Core
Orchestration_Core -- "orchestrates actions and assigns tasks to" --> Agent_System
Orchestration_Core -- "manages execution and state of tasks via" --> Task_Execution_Layer
Agent_System -- "interacts with for reasoning and tool execution" --> LLM_Tool_Integration
Agent_System -- "accesses and updates context from" --> Memory_Knowledge_Base
LLM_Tool_Integration -- "provides responses and tool execution results back to" --> Agent_System
Memory_Knowledge_Base -- "provides retrieved context and knowledge to" --> Agent_System
Orchestration_Core -- "emits events to" --> Observability_Telemetry
Agent_System -- "emits events to" --> Observability_Telemetry
Task_Execution_Layer -- "emits events to" --> Observability_Telemetry
LLM_Tool_Integration -- "emits events to" --> Observability_Telemetry
Memory_Knowledge_Base -- "emits events to" --> Observability_Telemetry
Observability_Telemetry -- "provides insights/logs to" --> CLI_Deployment
click CLI_Deployment href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/crewAI/CLI_Deployment.md" "Details"
click Orchestration_Core href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/crewAI/Orchestration_Core.md" "Details"
click Agent_System href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/crewAI/Agent_System.md" "Details"
click Task_Execution_Layer href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/crewAI/Task_Execution_Layer.md" "Details"
click LLM_Tool_Integration href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/crewAI/LLM_Tool_Integration.md" "Details"
click Memory_Knowledge_Base href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/crewAI/Memory_Knowledge_Base.md" "Details"
click Observability_Telemetry href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/crewAI/Observability_Telemetry.md" "Details"
The crewAI architecture is designed as an intelligent AI Agent Orchestration Framework, centered around the Orchestration Core which manages the execution of multi-agent workflows. User interaction begins with the CLI & Deployment component, initiating a Crew or Flow that the Orchestration Core then manages. Within this orchestration, individual Agent System instances perform tasks, leveraging the LLM & Tool Integration layer for their reasoning and external interactions, and drawing upon the Memory & Knowledge Base for contextual awareness and information retrieval. The Task Execution Layer ensures efficient task management throughout the workflow. All critical activities across these components are monitored by the Observability & Telemetry system, providing comprehensive insights into the framework's operation. This modular design facilitates clear data and control flow, making it ideal for visual representation in a flow graph diagram.
CLI & Deployment [Expand]
The user's primary interface for initiating, configuring, and deploying AI crews and flows.
Related Classes/Methods:
Orchestration Core [Expand]
The central workflow engine responsible for defining, managing, and executing multi-agent crews and complex flows.
Related Classes/Methods:
Agent System [Expand]
Encapsulates the intelligence, roles, and decision-making logic of individual AI agents within a crew.
Related Classes/Methods:
Task Execution Layer [Expand]
Manages the lifecycle, execution, and output handling of individual tasks assigned to agents.
Related Classes/Methods:
LLM & Tool Integration [Expand]
Provides a unified interface for agents to interact with Large Language Models and external tools for reasoning and action.
Related Classes/Methods:
Memory & Knowledge Base [Expand]
Manages all forms of information storage and retrieval, including contextual memory and RAG-based knowledge access for agents and crews.
Related Classes/Methods:
Observability & Telemetry [Expand]
A cross-cutting component for capturing and emitting detailed execution events for monitoring, debugging, and performance analysis.
Related Classes/Methods: