red2video

Convert Reddit posts into AI-generated short-form videos. Paste a Reddit URL and get a ready-to-upload MP4 with AI voiceover, generated images, and animated captions — in minutes.

How It Works

1. Fetch — Paste any Reddit post URL to extract the post and top comments
2. Script — Claude AI converts the content into a structured video script with scene-by-scene narration and image prompts
3. Characters — AI extracts character descriptions for visual consistency across scenes
4. Audio — OpenAI TTS generates voiceover for each scene (6 voice options)
5. Images — DALL-E 3 or SeeDream generates an image per scene
6. Render — Remotion renders everything into an MP4 with animated title card, scene transitions, and outro
7. Metadata — Claude generates YouTube-optimized titles, description, and tags

Supports 16:9 (YouTube), 9:16 (Shorts/TikTok/Reels), and 1:1 (Instagram) aspect ratios.

Prerequisites

• Node.js 18+
• FFmpeg — must be installed and available in your PATH
  • macOS: brew install ffmpeg
  • Windows: ffmpeg.org/download
  • Linux: sudo apt install ffmpeg
• API accounts (see Environment Variables)

Quick Start

git clone https://github.com/Didier187/red2video.git
cd red2video
cp .env.example .env   # fill in your API keys
npm install
npm run dev            # starts on http://localhost:3000

Environment Variables

Copy .env.example to .env and fill in your keys:

Variable	Required	Description
ANTHROPIC_API_KEY	Yes	Anthropic Console — used for script generation, character extraction, prompt enhancement, and YouTube metadata
OPENAI_API_KEY	Yes	OpenAI Platform — used for TTS audio and DALL-E 3 image generation
ARK_API_KEY	No	ByteDance ARK API — required only if you use the SeeDream image provider

Image Providers

• DALL-E 3 (default) — OpenAI's image generation, best quality, requires OPENAI_API_KEY
• SeeDream — ByteDance's image model via the ARK API, requires ARK_API_KEY

You can switch providers per-run in the UI.

Customizing the Outro Card

The outro card (channel name, social handle) is configured in .cta.json at the project root. This file is not committed — create it if you want a custom outro:

{
  "channelName": "Your Channel",
  "socialHandle": "@yourhandle"
}

If the file doesn't exist, the outro uses generic placeholder text.

Scripts

npm run dev       # Start dev server on port 3000
npm run build     # Build for production
npm run preview   # Preview production build
npm run test      # Run tests with Vitest
npm run check     # Prettier + ESLint fix

Storage

All generated media is stored locally in hidden directories at the project root:

.script-store/    # JSON script + character config per session
.media-store/     # Generated audio, images, and rendered video

These are gitignored and can be deleted at any time.

Architecture

See ARCHITECTURE.md for a full system diagram, data flow sequence, component breakdown, and API endpoint reference.

For AI agent coding guidelines, see AGENTS.md.

Contributing

See CONTRIBUTING.md.

License

MIT — see LICENSE.