Finda2aFinda2a
C

CrewAI Agent with A2A Protocol

Featured Service
Default
87 views

Introduction

This sample demonstrates a simple image generation agent built with CrewAI and exposed through the A2A protocol.

#CrewAI
#Agent Workflow
#LLM

Features

  • Text-to-image generation using Google Gemini
    Generate images from natural language descriptions powered by Gemini API.

  • Support for modifying existing images using references
    Enhance or edit images by providing reference inputs.

  • Robust error handling with automatic retries
    Automatically retries failed requests with built-in error management.

  • Optional file-based cache persistence
    Choose between in-memory or file-based caching for storing generated results.

  • Improved artifact ID extraction from queries
    Better parsing and handling of identifiers embedded in text queries.


Limitations

  • No true streaming
    CrewAI does not currently support real-time streaming responses.

  • Limited agent interactions
    No support for multi-turn conversations between agents.

Key Components

  • CrewAI Agent: Image generation agent with specialized tools
  • A2A Server: Provides standardized protocol for interacting with the agent
  • Image Generation: Uses Gemini API to create images from text descriptions
  • Cache System: Stores generated images for retrieval (in-memory or file-based)

Prerequisites

  • Python 3.12 or higher
  • UV package manager (recommended)
  • Google API Key (for Gemini access)

Sequence Diagram

CrewAI Diagram

CrewAI Diagram