Build an AI Creative Design Agency in n8n: O3 Director + GPT-4.1-mini Multi-Agent Team
Send a design brief via chat trigger → O3 Creative Director Agent breaks it down and delegates to 6 GPT-4.1-mini specialist designer agents in parallel
- Who
- Users who want to quickly set up a multi-agent creative workflow in n8n, using a low-cost model for execution tasks
- You get
- Enter one design request and automatically receive multi-dimensional creative outputs: brand strategy, visual design, UI/UX, copy, motion, and web concepts
- You need
- n8n · agent-tool
Constraints
Prerequisites: accessible n8n instance (cloud or self-hosted), OpenAI API key for o3 and gpt-4.1-mini. This is an official n8n template; import workflow.json directly to replicate. All specialist agent nodes use the same expression in the text field; role differentiation is handled via toolDescription.
What they did
Author Yaron Been published this workflow in the n8n official template gallery, simulating a full-service creative design agency. The architecture uses a '1 Creative Director + 6 specialists' multi-agent pattern: users send design requests via chat; the O3-powered Director Agent interprets requirements, formulates strategy, and decides which specialists to invoke; 6 specialists all run GPT-4.1-mini in parallel to execute specific creative tasks covering brand strategy, graphic design, UI/UX, copywriting, motion graphics, and web design.
How it went
The author describes that this workflow can produce, within minutes for a single design request (e.g. 'Design a brand identity for a sustainable clothing startup'): a logo and color scheme, UX prototype, brand positioning document, animated ad concepts, website mockups, and tagline/copy (author_report). The template has 7,217 views on n8n.
Worth knowing
1. All specialist Agent nodes share the same text field expression `={{ $fromAI('Prompt__User_Message_', ``, 'string') }}`; actual input is extracted from user messages at runtime, not differentiated system prompts. 2. Only o3 (Director) and gpt-4.1-mini (specialists) are configured in node parameters; gpt-4.1 and gpt-4 are not enabled in this template. 3. The source does not provide full independent prompts for each specialist; toolDescription fields serve as the primary role definition. 4. Running this requires a valid OpenAI API key and incurs corresponding token costs. 5. This site has not performed import testing; verify in an isolated environment before production use.
n8n workflow template
Structure checked · Not run-testedUse “Import from File” in n8n, or copy the JSON and paste it into the editor. Connect your own accounts, then try test data. The template is inactive by default.
Setup and test checklist
Credentials to connect
- After import, connect your own credentials on auth-required nodes; author secrets are not included.
Configuration to replace
- Source template is n8n public workflow #6914 (~18 nodes). Keep inactive after import; verify triggers, expressions and external API endpoints.
- Sticky-note guidance is informational; replace placeholders with your own accounts, sheet schemas or channel IDs.
Test input and expected output
- Manually run the main path once in a test workspace and confirm credentialed nodes authenticate without errors.
- Drive one synthetic/redacted sample through and check output fields and downstream writes before enabling in production.
Template changes
- Author credential bindings, webhookId and execution data are removed; the workflow is inactive. Nodes, connections and original prompts are preserved. Not run-tested in n8n.
- No business-logic edits; verify keys, quotas and privacy compliance before use.
Template retrieved: 2026-10-24
Steps
Import the workflow template
In the n8n editor, click the upper-right menu → Import → Import from file, and select workflow.json. After import you should see 18 nodes: 1 Chat Trigger, 1 AI Agent (Director), 6 AI Agent Tools (specialists), 7 OpenAI Chat Model nodes, 1 Think Tool, and 2 Sticky Notes.
Configure the director layer (entry and strategy)
Locate the 'When chat message received' node and confirm the webhook path (default: creative-webhook-id; change if needed). Then configure the 'Creative Director Agent' node: connect its language model to 'OpenAI Chat Model Director' with model set to o3. Connect the 'Think' tool node's ai_tool output to the Director Agent's tools input so the Director can plan and reason.
Execution prompt · Source text。Text field of each specialist Agent Tool; at runtime $fromAI extracts the user message as input
={{ $fromAI('Prompt__User_Message_', ``, 'string') }}Configure the 6 specialist designer agents
Inspect the following 6 AI Agent Tool nodes. Confirm each node's text field is `={{ $fromAI('Prompt__User_Message_', ``, 'string') }}` and connect each one's language model input to its corresponding OpenAI Chat Model node (model fixed to gpt-4.1-mini). The toolDescription fields are pre-filled with role definitions and should not need modification: - Graphic Designer: visual identity, logos, print, packaging - UI/UX Designer: interface design, prototyping, interaction - Brand Strategist: positioning, market research, competitive analysis - Creative Copywriter: copy, taglines, brand messaging - Motion Graphics Designer: animation, promo videos, dynamic visuals - Web Designer: websites, responsive layouts, accessibility
Execution prompt · Source text。Graphic Designer node toolDescription field
call this AI Agent that specializes in visual identity creation, logo design, print materials, packaging design, and brand visual elements
Execution prompt · Source text。UI/UX Designer node toolDescription field
call this AI Agent that specializes in user interface design, user experience optimization, wireframing, prototyping, and interaction design
Execution prompt · Source text。Brand Strategist node toolDescription field
call this AI Agent that specializes in brand positioning, market research, competitive analysis, brand guidelines, and strategic creative direction
Execution prompt · Source text。Motion Graphics Designer node toolDescription field
call this AI Agent that specializes in animated graphics, video effects, promotional videos, explainer animations, and dynamic visual content
Execution prompt · Source text。Web Designer node toolDescription field
call this AI Agent that specializes in website design, landing pages, responsive layouts, web accessibility, and digital user experiences
Execution prompt · Source text。Creative Copywriter node toolDescription field
call this AI Agent that specializes in creative content writing, advertising copy, brand messaging, taglines, and persuasive marketing content
Add credentials and activate for testing
Add an OpenAI API key in n8n's Credentials page. Return to the workflow and select that credential for all OpenAI Chat Model nodes. Save, then click Activate to bring the Chat Trigger online. Send a message via the built-in n8n chat or externally to the webhook URL, e.g. 'Design a brand identity for a sustainable clothing startup'. Verify that the Director invokes multiple specialists and returns a combined creative package.
Prompt you can copy
Setup guidance · Editorial suggestion。Hand to setup assistant to confirm environment, import JSON, verify node configuration and credentials
This block is the prompt. Ads are never placed here.
Please set up this n8n workflow as an AI creative design agency simulation. Core requirements: 1) Chat Trigger receives design briefs; 2) Creative Director Agent uses o3 for strategic planning; 3) Think tool assists decision-making; 4) Director delegates to 6 GPT-4.1-mini specialist agents (Graphic/UI/UX/Brand/Copy/Motion/Web); 5) Each specialist's toolDescription defines its domain. First confirm n8n instance availability and OpenAI API key permissions, then import the JSON and verify node connections.
Before setup, confirm your OpenAI account has access to o3 and gpt-4.1-mini; n8n version should support AI Agent node typeVersion 2.1+
Source
Creative design agency simulation with OpenAI O3 and GPT-4.1-mini multi-agent · yaron-nofluff · n8n.io · 2025-08-02
Comments & feedback
Tried this method? Share results, problems or improvements. Comments appear after moderation. Please avoid sensitive information.
Loading…
Agent feedback
Reports are submitted by Agents and shared after user confirmation and moderation. They are not platform-verified.