English | 简体中文 | 繁體中文 | 한국어 | Deutsch | Español | Français | Italiano | Dansk | 日本語 | Polski | Русский | Bosanski | العربية | Norsk | Português (Brasil) | ไทย | Türkçe | Українська | বাংলা | Ελληνικά | Tiếng Việt | हिन्दी

AutoVio

Open-source AI video generation pipeline for SaaS teams and developers.
From a text prompt to a finished video — scenario, images, clips, editing, export.

📖 Docs · 🚀 Quick Start · 📡 API · 🤖 MCP Server

What is AutoVio?

Most AI tools handle one step of video creation. AutoVio handles the whole thing.

You describe what you want — a product, an idea, a story. AutoVio writes the scene-by-scene scenario, generates an image for each scene, animates those images into video clips, and assembles everything in a timeline editor. You export a finished MP4. Especially useful for SaaS product demos, feature announcements, and marketing videos.

The entire pipeline runs on your own infrastructure. You bring your own API keys. You own the output.

Text prompt  →  Scenario (LLM)  →  Images (Gemini / DALL-E)  →  Video clips (Veo / Runway)  →  Edit  →  Export

The Pipeline

AutoVio breaks video production into five steps that mirror how a human team would work:

Two generation modes:

• Style Transfer — Replicate the visual style of an existing video on new content
• Content Remix — Build from scratch using a project style guide and your prompts

Key Features

• Full end-to-end pipeline — one system from idea to exported MP4
• Multi-provider AI — mix and match LLMs, image models, and video models per project
• Reference video analysis — vision AI decodes style, tempo, and composition from any video
• Project style guides — lock in brand voice, color palette, camera style, and tone once; apply across all videos
• Asset library — upload product photos, logos, or screenshots; use them directly in videos or as style references
• Timeline editor — text overlays, image overlays, transitions, audio mixing, frame-accurate trimming
• Template system — save overlay compositions as reusable templates across works
• Resolution control — Portrait 9:16, Landscape 16:9, or Square 1:1; each provider gets the right format automatically
• REST API + OpenAPI — every feature is accessible programmatically
• MCP server — use AutoVio from Claude Code, Cursor, Claude Desktop, or any MCP client
• Self-hosted — runs on your machine or your server; no data leaves without your API keys

AI Providers

AutoVio is provider-agnostic. Configure different providers for each role:

New providers can be added by implementing the IImageProvider or IVideoProvider interface.

Use Cases

Developers & AI Coding Assistants

AutoVio has a full MCP server. Your AI coding assistant can generate product demo videos without leaving the editor:

• Claude Code — run autovio_works_create after shipping a feature
• Cursor — generate tutorial videos for code changes inline
• Claude Desktop — describe a video in conversation, have it built automatically

Automation Workflows

The REST API connects to any automation platform:

• n8n / Make / Zapier — trigger video generation from webhooks, CRM events, or schedules
• CI/CD pipelines — auto-generate release announcement videos on every deploy
• Content calendars — batch-produce social media videos from a content schedule

SaaS & Product Teams

• Turn feature specs into product demo videos
• Generate localized video variants from a single scenario
• Create onboarding videos from documentation
• Automate release announcement videos for every new SaaS feature
• Maintain brand consistency across all video output with style guides

Researchers & Builders

• Experiment with new AI video providers without rebuilding infrastructure
• Use the REST API as a backend for your own video product
• Extend the pipeline with custom providers, prompts, or export formats

Quick Start

Requirements

• Bun >= 1.0 (or Node.js >= 18)
• MongoDB — local or Atlas
• FFmpeg — for video export (brew install ffmpeg / apt install ffmpeg)
• At least one AI provider API key (Google Gemini is free to start)

1. Clone and install

git clone https://github.com/Auto-Vio/autovio.git
cd autovio
bun install

2. Configure

cp .env.example .env
# Open .env and set MONGODB_URI and JWT_SECRET

3. Start

bun run dev

• Frontend: http://localhost:5173
• Backend API: http://localhost:3001
• OpenAPI docs: http://localhost:3001/api/docs

MCP Server

The autovio-mcp package ships a full MCP server with 25+ tools covering the entire AutoVio API. Connect it to Claude Code, Claude Desktop, Cursor, or any MCP-compatible client and generate videos through conversation.

Claude Code:

claude mcp add autovio-mcp -- npx -y autovio-mcp \
  --autovio-base-url http://localhost:3001 \
  --autovio-api-token YOUR_TOKEN \
  --llm-model gemini-2.5-flash \
  --llm-api-key YOUR_KEY \
  --image-model gemini-2.5-flash-image \
  --image-api-key YOUR_KEY \
  --video-model veo-3.0-generate-001 \
  --video-api-key YOUR_KEY

Claude Desktop / Cursor (claude_desktop_config.json):

{
  "mcpServers": {
    "autovio": {
      "command": "npx",
      "args": [
        "-y", "autovio-mcp",
        "--autovio-base-url", "http://localhost:3001",
        "--autovio-api-token", "YOUR_TOKEN",
        "--llm-model", "gemini-2.5-flash",
        "--llm-api-key", "YOUR_KEY",
        "--image-model", "gemini-2.5-flash-image",
        "--image-api-key", "YOUR_KEY",
        "--video-model", "veo-3.0-generate-001",
        "--video-api-key", "YOUR_KEY"
      ]
    }
  }
}

See the MCP documentation for the full setup guide and tool reference.

Project Structure

AutoVio/
├── packages/
│   ├── backend/     # Express API — routes, AI providers, FFmpeg export
│   ├── frontend/    # React + Vite — 5-step pipeline UI, timeline editor
│   └── shared/      # TypeScript types shared between packages
└── package.json     # Bun/npm workspace root

Contributing

AutoVio is at an early stage and actively evolving. Contributions are welcome in any form:

• Bug reports — open an issue with reproduction steps
• New AI providers — implement IImageProvider or IVideoProvider and open a PR
• UI improvements — the frontend is React + TailwindCSS + Zustand
• Documentation — the docs site lives in AutoVio-Docs
• Ideas and feedback — open a discussion or issue

To get started, read the documentation and explore the codebase. The provider interfaces in packages/backend/src/providers/interfaces.ts are a good entry point for adding new AI integrations.

Repositories

Scripts

License

AutoVio is licensed under PolyForm Noncommercial 1.0.0.

Free for personal, educational, and non-commercial use. For commercial use, contact the maintainers to discuss licensing.