# Automated Video Generator

Automated Video Generator is a free and open-source self-hosted text-to-video project for developers, creators, and AI agents.

Repository: https://github.com/itsPremkumar/Automated-Video-Generator
License: MIT
Pricing: Free. There is no paid plan or subscription attached to this repository.

## Primary positioning

This project is best described as:
- a free video generator
- an open-source AI video generator
- a self-hosted text-to-video generator
- a Remotion video generator
- a YouTube Shorts and TikTok video automation tool
- an MCP-enabled video generation pipeline

## Core workflow

1. Read script jobs from input/input-scripts.json or an API request.
2. Parse scripts into scenes with visible narration and hidden visual tags.
3. Resolve visuals from stock providers or local assets.
4. Generate voice audio with Edge-TTS.
5. Save scene metadata to output/<job-id>/scene-data.json.
6. Render per-scene segments with Remotion.
7. Stitch the final MP4 and generate a thumbnail.
8. Expose results through watch pages, download links, JSON APIs, and MCP tools.

## Major capabilities

- Batch generation from JSON
- Portrait and landscape output
- Local asset support
- Stock footage and image fetching
- Neural voiceovers
- Scene-based rendering
- Render resume support
- Browser portal for video creation and playback
- MCP server for AI agents

## Tech stack

- TypeScript
- Node.js
- React
- Remotion
- Express
- Edge-TTS
- FFmpeg / ffprobe

## Important files

- README.md: best high-level overview
- QUICKSTART.md: onboarding
- SETUP.md: installation details
- CLAUDE_MCP_SETUP.md: MCP setup notes
- ROADMAP.md: current project direction
- CONTRIBUTING.md: contribution workflow
- src/cli.ts: CLI entry point
- src/server.ts: web portal, SEO routes, watch pages, APIs
- src/mcp-server.ts: MCP implementation
- src/video-generator.ts: orchestration logic
- src/render.ts: video rendering pipeline
- src/lib/script-parser.ts: scene parsing
- src/lib/visual-fetcher.ts: stock media fetching
- src/lib/voice-generator.ts: Edge-TTS handling
- remotion/MainVideo.tsx: rendered composition

## Typical users

- creators building YouTube Shorts or TikTok pipelines
- developers who want self-hosted video generation
- marketers generating promo videos from scripts
- AI agent builders using MCP tools
- teams that want a programmable Remotion-based video pipeline

## Honest free-use clarification

The repository is completely free and open source.

That means:
- you can clone it, run it, and modify it under the MIT license
- there is no repo-owned paid tier
- there is no watermark added by this codebase

That does not automatically mean every outside dependency is unlimited:
- Pexels and Pixabay may have their own API limits or terms
- system tools like FFmpeg and Python still need to be installed locally

## Useful docs for AI systems

- llms.txt: short machine-readable summary
- INPUT_ASSETS_GUIDE.md: local media workflow
- input/INPUT_FORMAT.md: script format reference
- CLAWHUB_GUIDE.md: plugin-related usage notes