In short
The Drever project combines MDX, AI, and animation to create interactive presentations. We’ll explore why an “AI-first” approach in a slide-making tool isn’t just marketing—it’s a paradigm shift.
Presentations have long been stuck in the “slides + bullet points” paradigm. The Drever project offers a different approach: presentations as MDX code, generated by AI and featuring expressive animations.
The main focus is on readability and control. MDX lets you write slides as text rather than dragging and dropping blocks in a visual editor. This provides version control, component reusability, and proper Git diffs. For engineering teams where presentations are part of technical communication, this is a strong argument in and of itself.
“AI-first” here isn’t just a marketing label. The repository structure speaks for itself: the .agents/ and .claude-plugin/ directories imply that content generation and editing are handled by agents, not through manual input. The presentation becomes an artifact that can be generated, modified, and maintained programmatically.
What’s important from a practical standpoint: the project is young (2 stars, 0 forks at the time of this review), but the structure is well-thought-out—end-to-end tests, a plugin-based architecture, and documentation. This isn’t a weekend hack, but a thoughtful attempt to solve a problem. However, there are still few ready-made examples and details about how AI agents work available in the open, so the real value of implementation will have to be tested hands-on.