Summary

StoryGen-Atelier is an AI-assisted tool designed for generating storyboards and full videos. It targets creators and developers seeking to automate video production workflows, leveraging Google's Gemini for script and frame generation, and Vertex AI Veo for video synthesis, ultimately streamlining the creation of visual narratives.

How It Works

The project employs an "Interpolation Chain" strategy for video generation. It begins by using Gemini to analyze pairs of adjacent storyboard shots, generating detailed transition prompts and suggested durations. Subsequently, Vertex AI's Veo model synthesizes video clips based on these prompts, the starting shot, and the ending shot. A final "Closing Shot" clip is also generated. FFmpeg then performs a lossless stitch of all generated clips using the concat protocol and stream copy (-c copy) to assemble the complete video efficiently.

Quick Start & Requirements

• Primary Install/Run: Execute chmod +x start_servers.sh once, then run ./start_servers.sh from the root directory. This automatically starts the backend (port 3005) and frontend (port 5180).
• Prerequisites: Node.js 18+, npm, ffmpeg, and a Google Cloud Project with Vertex AI API enabled. A Gemini API Key is also required.
• Configuration: Environment variables, including API keys and project details, must be configured in backend/.env (copy from .env.example).
• Documentation: Links to example storyboard styles are provided.

Highlighted Details

• Storyboard Generation: Utilizes Gemini for generating text scripts and image frames, supporting custom styles and shot counts.
• Video Synthesis: Leverages Vertex AI Veo for creating intermediate transition clips, driven by AI-analyzed storyboard transitions.
• Efficient Assembly: Employs FFmpeg with stream copying (-c copy) for lossless stitching of generated video clips.
• Management Features: Includes a built-in logs dashboard (SQLite persistence) and a gallery for managing storyboards and videos.

Maintenance & Community

No specific details regarding contributors, sponsorships, community channels (e.g., Discord, Slack), or a public roadmap are present in the provided README.

Licensing & Compatibility

The README does not specify a software license. Compatibility for commercial use or closed-source linking is therefore undetermined.

Limitations & Caveats

The project is dependent on Google Cloud services, requiring active GCP project setup and associated costs. The use of preview models for Gemini and Vertex AI Veo suggests the system may be in an experimental or pre-release phase, potentially subject to changes or instability.