I built SlideBot AI — an open-source AI-powered presentation generator that creates professional slides from natural language input.

Why I built this:

At my company, we spend hours every week creating presentation decks. The existing AI tools either generate generic-looking slides or require too much manual tweaking. I wanted something that:

1. Takes a topic/outline and generates a complete presentation 2. Lets you iterate through natural conversation ("make slide 3 more concise", "add a risk section") 3. Supports voice input — upload a meeting recording and it extracts key points automatically 4. Maintains consistent brand styling across all slides

How it works:

- Input your idea (text, outline, or audio recording) - AI generates an outline → you refine it through conversation - AI creates a design scheme → you refine it through conversation - AI generates each slide as a high-quality image - Download as a ZIP package

Tech stack: React + FastAPI + Google Gemini API

What makes it different from other AI slide tools:

- Fully open-source (MIT license) - Conversational iteration at every step — not just one-shot generation - Voice-to-slides pipeline (meeting recording → presentation) - Born from actual daily business use, not a hackathon project

This started as an internal tool and has been used in production for several weeks. The iterative refinement workflow was the key insight — the first draft is never perfect, but being able to say "add more data to slide 5" or "make the conclusion stronger" makes a huge difference.

Would love to hear your feedback. What features would make this more useful for your workflow?

GitHub: https://github.com/tonyqinatcmu/SlideBot-AI/blob/main/README...