IdeaMade is now live — turning app ideas into editable Blueprints before building
A few weeks ago, I started building IdeaMade around a simple question:
Can AI help people define what they actually want to build before generating the app?
After several iterations, the first public version is now available.
The current experience:
Describe your app idea → Generate an initial Blueprint → Confirm key decisions → Update the Blueprint
Instead of asking users to complete a long requirements form, IdeaMade first creates a Blueprint from their idea.
Then AI identifies up to 3 important decisions that are still unclear and may affect the product direction.
For example:
Should bookings be automatically confirmed or manually approved?
Which user roles need different permissions?
Which workflow is the core experience?
Users can answer these questions or use AI recommendations, and the Blueprint updates on the same page.
The current Blueprint includes:
Product goal
Users and roles
Key workflows
Main pages
Important assumptions
The first public version is free to try:
A few things I’m trying to learn:
Are these 3 questions actually the decisions users need help with?
Does seeing the Blueprint make the product idea clearer?
Is this a better starting point than immediately generating an app?
This is still an early experiment. I’d love feedback from people who have built products with AI tools.
Try it with a real idea and tell me:
What did the Blueprint get wrong, or what important decision did it miss?
Of your three questions, the second ("are these 3 the decisions users need help with") is the only one that really matters, the others follow from it. Sharp questions make the Blueprint clarifying and beat generating an app cold. Generic ones, nothing lands. Obsess over question quality.
A test for a good question: it references something specific the user said, not something generic about the app type. "Should bookings auto-confirm" is fine but templatey, any booking app gets it. "You described walk-ins AND reservations, which wins when they conflict for a table" is the jolt, because it proves the AI read their idea, not its category.
Reviewing sessions, tag each of the 3 as generic vs idea-specific. That ratio is your product quality.