automated-video-generator

v1.0.0

Published

2 months ago

Free and open-source self-hosted AI text-to-video generator built with Remotion, Edge-TTS, stock media fetching, batch rendering, a local web portal, and MCP integration.

Automated Video Generator

Automated Video Generator is a self-hosted text-to-video pipeline for developers, creators, and AI agents. Give it a script and it can fetch visuals, generate voiceovers, render scenes with Remotion, and export a ready-to-share video.

If you are searching for a free video generator, open-source AI video generator, Remotion video generator, YouTube Shorts generator, TikTok video generator, or self-hosted text-to-video tool, this repo is built for that workflow.

This is not a fake "free trial" generator. The project itself is completely free and MIT-licensed. There is no paid plan, no subscription, and no watermark added by this codebase. Optional third-party services such as Pexels or Pixabay may still have their own quotas or terms.

Why this repo gets attention

Free and open source under the MIT license
Self-hosted video generation you control locally
Text-to-video pipeline with Remotion and React
Edge-TTS voiceovers with multiple neural voice options
Stock media fetching from Pexels and Pixabay
Local asset support for your own images and videos
Batch generation for multiple videos in one run
Local web portal for generating, previewing, and sharing videos
Built-in MCP server for Claude Desktop, Claude Code, and other MCP clients

Best use cases

YouTube Shorts automation
TikTok and Reels content pipelines
Faceless YouTube channels
Marketing videos and product promos
Explainer videos and tutorials
Programmatic content generation for AI agents

Core features

Script-driven video generation from plain text or JSON
Director mode with [Visual: ...] tags for exact visual control
Automatic scene parsing and timeline generation
Neural voice generation using edge-tts
Portrait and landscape video output
Resumable segmented rendering with Remotion
Render thumbnails for completed videos
Browser portal for generation, status tracking, playback, and downloads
MCP tool interface for agentic workflows

Quick start

Prerequisites

Node.js 18+
npm
Python 3.8+
FFmpeg available on your PATH

Install

git clone https://github.com/itsPremkumar/Automated-Video-Generator.git
cd Automated-Video-Generator
npm install
pip install -r requirements.txt

Configure environment variables

Copy .env.example to .env and add your keys.

# Free stock media API keys
PEXELS_API_KEY=your_key_here
PIXABAY_API_KEY=your_key_here

# Optional but recommended for public deployments
PUBLIC_BASE_URL=https://your-domain.example

# Optional defaults
PORT=3001
VIDEO_ORIENTATION=portrait
VIDEO_VOICE=en-US-GuyNeural

PEXELS_API_KEY is the main one to start with, and Pexels offers a free API key.

Generate a video

Create input/input-scripts.json with one or more jobs:

[
  {
    "id": "youtube-shorts-demo",
    "title": "3 Productivity Habits That Actually Work",
    "orientation": "portrait",
    "voice": "en-US-JennyNeural",
    "script": "[Visual: focused person at desk] Want better focus every day? [Visual: checklist and timer] Try these three simple habits."
  }
]

Run the pipeline:

npm run generate

The final video will be written to output/<id>/.

Local web portal

Run the local portal:

npm run dev

Then open:

http://localhost:3001/

The portal lets you:

Start a render from the browser
Track progress on a job page
Watch completed videos
Download the final MP4
Expose SEO-ready pages if you deploy it publicly

Remotion studio

Preview templates and compositions locally:

npm run remotion:studio

MCP and AI agent support

This project ships with an MCP server, so AI tools can create and manage videos through chat-driven workflows.

Start the MCP server:

npm run mcp

Useful for:

Claude Desktop
Claude Code
Other Model Context Protocol clients

Project health

CI runs on pushes and pull requests
Dependabot keeps npm and GitHub Actions dependencies fresh
Issue templates make bug reports and feature requests easier to review
A pull request template helps contributors ship cleaner changes

How it works

Parse a script into scenes and durations.
Fetch stock visuals or use local assets from input/input-assests/.
Generate voiceover audio with Edge-TTS.
Save scene data into output/<job-id>/scene-data.json.
Render scene segments with Remotion.
Stitch the final MP4 and thumbnail for playback and sharing.

Available commands

npm run generate         # Generate videos from input/input-scripts.json
npm run resume           # Resume an interrupted generation run
npm run segment          # Rebuild from existing scene data
npm run remotion:studio  # Open Remotion studio
npm run remotion:render  # Render using the render pipeline
npm run dev              # Start the local web portal
npm run mcp              # Start the MCP server
npm run typecheck        # Validate TypeScript before opening a PR

Project structure

src/
  cli.ts                 Batch generation entry point
  server.ts              Local web portal and API
  mcp-server.ts          MCP server entry point
  video-generator.ts     Generation pipeline
  render.ts              Segmented Remotion renderer
  lib/
    script-parser.ts
    visual-fetcher.ts
    voice-generator.ts
    cleaner.ts
remotion/
  MainVideo.tsx
  SingleSceneVideo.tsx
  Root.tsx
input/
  input-scripts.json
  input-assests/         Local images and videos
output/                  Generated videos
public/                  Job assets served by the portal

FAQ

Is this a completely free video generator?

Yes. The project itself is completely free and open source under the MIT license. There is no paid plan attached to this repo. Optional external services may have their own rules or usage limits.

Is this an open-source AI video generator?

Yes. It is an open-source text-to-video pipeline that uses AI voice generation plus deterministic media selection and Remotion rendering.

Can I use this for YouTube Shorts, TikTok, and Reels?

Yes. Use portrait for 9:16 output and landscape for 16:9 videos.

Does this add a watermark?

No watermark is added by this project.

Can I use my own images and videos?

Yes. Put files in input/input-assests/ and reference them with [Visual: filename.mp4] or [Visual: filename.jpg].

Can I self-host it?

Yes. You can run it locally, in Docker, or behind your own deployment setup.

GEO and AI-friendly docs

These files make the project easier for AI tools and answer engines to understand:

Roadmap and contributing

Growth assets

Contributing

Issues, feature requests, docs improvements, and pull requests are welcome.

If this repo helps you, please star it on GitHub:

https://github.com/itsPremkumar/Automated-Video-Generator