Info To Video — AI skill for Claude Code

Infographic to Video

A Claude Code skill that converts **infographics, articles, or video scripts** into narrated MP4 videos with AI voiceover and burned-in subtitles.

What This Does

You give Claude an infographic image, an article, or a video script, and it automatically:

1. Reads your input — extracts content from images, text, or URLs
2. Writes a narration script — conversational, educational voiceover (not just reading the text)
3. Generates HTML slides — using frontend-slides for stunning visuals
4. Produces AI voiceover — via Edge TTS (free, no API key needed)
5. Burns in subtitles — word-level timing, rendered with Pillow
6. Assembles the final MP4 — with ffmpeg

The entire pipeline runs locally. No cloud APIs, no costs.

Demo

Input: an infographic about options trading strategies

Output: a 4-minute narrated video with slides, voiceover, and subtitles

Installation

git clone https://github.com/ftcalvinwang-ai/info-to-video.git ~/.claude/skills/info-to-video

Then use it by typing `/info-to-video` in Claude Code.

Requirements

• Claude Code CLI
• Python 3.9+ with edge-tts and Pillow
• Node.js (for Playwright screenshots)
• ffmpeg
• For YouTube input: youtube-transcript-api, yt-dlp, and optionally openai-whisper

Install dependencies

pip3 install edge-tts Pillow youtube-transcript-api yt-dlp
pip3 install openai-whisper  # optional, for YouTube videos without subtitles
brew install ffmpeg  # macOS

Playwright is installed automatically on first run (via npm in a temp directory).

Usage

Via Claude Code (Recommended)

/info-to-video

Then follow the prompts — share your infographic, confirm the script, choose a voice, and get your video.

Standalone Script

You can also use `generate-video.py` directly if you already have HTML slides and a narration JSON:

python3 genera