Building Two SaaS MVPs with AI: What I Learned

Description and use

Two AI-powered SaaS MVPs were built using a shared tech stack and Docker Compose for rapid development.

Problem

Create functional SaaS MVPs quickly using AI tools and a consistent tech stack.

Outcome

Both AI Document Platform and AI Code Evaluation Platform MVPs are operational, demonstrating rapid MVP development using a unified stack and Docker-based automation.

Workflow

• 1. Define product idea and required features.
• 2. Choose a common tech stack (Python/FastAPI, React/TS, Vite, PostgreSQL, Redis).
• 3. Set up Docker and Docker Compose for development and deployment.
• 4. Implement backend services with FastAPI and Python.
• 5. Build frontend with React, TypeScript, and Vite.
• 6. Configure PostgreSQL for data persistence and Redis for caching.
• 7. Use OpenAI or local models to generate embeddings for the AI Document Platform.
• 8. Create evaluation prompts and tests for the AI Code Evaluation Platform.
• 9. Automate setup with Make and Docker Compose scripts.
• 10. Run "docker compose up" to launch both services.

Evidence

Both projects use a similar stack: Backend: Python and FastAPI, Frontend: React, TypeScript, and Vite, Database: PostgreSQL, Caching: Redis, Development and deployment: Docker and Docker Compose

Source

The projects aren’t perfect, and there’s still room to improve them, but both are up and running.

Source

You’ll need Docker and Make installed. I’ve automated much of the setup, so you can try starting the projects with: bash docker compose up

Source