slides-grab addresses the gap in AI-powered slide generation tools by providing a visual, browser-based editor for HTML/CSS slides. It enables users to directly select and edit content areas within AI-generated presentations, facilitating iterative refinement. Targeted at users leveraging AI for slide creation and developers integrating AI into presentation workflows, it offers a more interactive and efficient alternative to purely generative approaches.

How It Works

The project generates slides as self-contained HTML and CSS files. A core component is a JavaScript-based editor allowing users to visually select areas (using bounding boxes) and perform direct text edits. It integrates with AI agents like Claude Code and Codex, enabling agent-powered rewrites and modifications. The approach prioritizes an interactive editing loop, allowing for immediate visual feedback on AI-driven changes, unlike purely code-based generation.

Quick Start & Requirements

Installation is straightforward via npm: npm install slides-grab followed by npx playwright install chromium. Alternatively, clone the repository for development: git clone https://github.com/vkehfdl1/slides-grab.git && cd slides-grab && npm ci && npx playwright install chromium. Requires Node.js version 18 or higher. Detailed installation guides for Claude Code and Codex are available here and here.

Highlighted Details

• Interactive editor with bounding box selection and direct text manipulation.
• AI agent integration for in-place slide content editing and rewriting.
• Export capabilities to PDF (capture or print modes, high-resolution options) and experimental PPTX/Figma formats.
• Strict image handling policy: local images in ./assets/ preferred, remote URLs warned, absolute paths unsupported.
• slides-grab validate command ensures slide HTML integrity and asset path correctness.
• Tldraw integration for embedding generated diagrams as SVG assets.

Maintenance & Community

The project acknowledges inspiration from "ppt_team_agent by Builder Josh." No specific community channels (Discord, Slack), roadmap links, or prominent maintainer/sponsor information are provided in the README.

Licensing & Compatibility

The project is released under the MIT license. This permissive license generally allows for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

PPTX and Figma export functionalities are explicitly labeled as experimental and unstable, potentially requiring manual post-export cleanup. Remote HTTPS images are supported with warnings, and absolute filesystem paths for images are unsupported. Legacy Tldraw files require resaving to function correctly.