This is a submission for the Notion MCP Challenge
What I Built
I built Merch OS, an AI-powered workflow for turning a merch brief into a structured execution workspace inside Notion.
The idea came from a real creative operations problem: merch and apparel workflows often get scattered across chats, notes, design files, and mental to-do lists. I wanted to build something that takes a brief and turns it into a system that is actually ready for execution.
With Merch OS, a user can:
- enter a merch brief in a frontend form
- generate multiple AI-driven design directions
- select the strongest direction
- automatically create a Notion workspace for the project
The generated Notion workspace includes:
- a Workspace Overview
- a Brand Brief
- a Design Directions page
- a Feedback page
- a Production Tasks database
- a Project Milestones database
The production task database is dynamic. It adapts based on the selected direction, garment types, typography style, print style, and design structure. The milestone database also uses the project start date and launch date to generate a simple timeline for execution.
So instead of AI output staying in a chat window, it becomes a live workspace inside Notion that can actually be used to run the project.
Video Demo
https://www.loom.com/share/fe849b96e4a54130bb091880e7393fd2
Show us the code
https://github.com/mshaikhly/merch-os
How I Used Notion MCP
Notion MCP is a core part of the workflow.
After the user fills in the brief and selects a direction, Merch OS uses Notion MCP to create the project workspace structure inside Notion. MCP is responsible for creating the main workspace and page-based project documentation, including:
- the parent project workspace page
- the brief page
- the design direction page
- the feedback page
This is what turns the workflow from “AI generated some ideas” into “AI created a usable project system inside Notion.”
For the database layer, I used the official Notion API alongside MCP because the MCP server surface I was working with exposed page operations cleanly, while modern Notion database creation is now handled through the newer database/data source model. So the final workflow is a hybrid:
- Notion MCP for creating and structuring the workspace
- Notion API for creating the task and milestone databases
- AI generation for producing the actual creative direction and execution logic
What this unlocks is a much more practical workflow:
- the user enters a brief
- AI generates creative directions
- the user selects a direction
- Notion MCP creates the workspace
- the system creates execution databases for production and milestones
That means the output is not just inspiration. It becomes a real workspace that can be edited, reviewed, and used by a team.
Notes
A big focus of this project was making the workflow feel practical rather than gimmicky. I wanted the AI to help with both ideation and operational structure, and Notion MCP was the key piece that made that possible.
Top comments (1)
Good