graph TD %%{init: {'theme': 'mc','layout': 'elk'}}%% ChatInput-g80eg[<div><img alt="logo" src="/_astro/messages-square.BaSDmT6g.svg" style="height: 20px !important;width: 20px !important"/></div>Chat Input] style ChatInput-g80eg stroke:#a170ff AdvancedAgent-sj22e[Agent] style AdvancedAgent-sj22e stroke:#a170ff GoogleGenerativeAIModel-t580h[Gemini] style GoogleGenerativeAIModel-t580h stroke:#a170ff GeminiImageGenerator-k3t74[Gemini Image Generator] style GeminiImageGenerator-k3t74 stroke:#a170ff ChatOutput-64ffn[<div><img alt="logo" src="/_astro/messages-square.BaSDmT6g.svg" style="height: 20px !important;width: 20px !important"/></div>Chat Output] style ChatOutput-64ffn stroke:#a170ff ChatInput-g80eg -.- AdvancedAgent-sj22e linkStyle 0 stroke:#a170ff GoogleGenerativeAIModel-t580h -.- AdvancedAgent-sj22e linkStyle 1 stroke:#a170ff GeminiImageGenerator-k3t74 -.- AdvancedAgent-sj22e linkStyle 2 stroke:#a170ff AdvancedAgent-sj22e -.- ChatOutput-64ffn linkStyle 3 stroke:#a170ff

📄 Inspirational Graphic Workflow Documentation

🧩 Overview

The Inspirational Graphic workflow enables an AI agent to transform ambiguous client briefs into concrete visual concepts. A user submits a message, the agent, powered by Gemini, analyzes the brief, proposes three divergent creative directions, and automatically generates supporting images with Gemini Image Generator. The final response is presented back to the user as a chat message.

⚙️ Main Features

User‑centric input collection with optional file attachments.
Conversational agent that crafts multiple creative directions based on a sophisticated system prompt.
Dynamic tool integration allowing the agent to generate images on demand.
Automated output rendering to a chat interface.
Configurable language model and image‑generation settings for flexibility.

🔄 Workflow Steps

Component Name	Role in the Workflow	Key Inputs	Key Outputs
Chat Input	Captures the user’s brief and any attached files.	User message (text), optional files, conversation metadata (IDs).	A structured `Message` containing the user’s input.
Advanced Agent	Interprets the brief, calls Gemini for text generation, and invokes Gemini Image Generator for visual assets.	User message (from Chat Input), Language model (Gemini), Tool (Gemini Image Generator), system prompt definition.	A conversational `Message` that includes text and references to any generated images.
Gemini (Language Model)	Produces the agent’s textual response using the prompt supplied by the agent.	Prompt text (derived from system prompt and user input), temperature, token limits.	Generated text wrapped in a `Message` ready for further processing.
Gemini Image Generator	Creates visual assets (logos, UI mock‑ups, branding elements) as requested by the agent.	Image prompt (text description), image model selection, aspect ratio, number of images, optional source files.	Binary image data that the agent incorporates into its response.
Chat Output	Renders the agent’s final message in the chat interface.	Response `Message` (text + image references), sender information.	Displayed chat message visible to the user.

🧠 Notes

The system prompt defines the agent’s role as a senior art director and sets clear expectations for creative output.
Memory and tool‑call limits are configured within the Advanced Agent but are not exposed in the documentation to keep the focus on functional flow.
The agent’s structured output schema is optional; if used, it enforces a JSON format for the response.
Streaming is disabled by default; enabling it would allow the agent’s thoughts to appear incrementally.
All interactions are stateless between runs unless a memory component is added; the workflow as presented is a single‑pass, end‑to‑end process.