character-series-generator

AI-powered tool that turns a one-line story idea into a complete animated series with consistent characters, voiceovers, and automated video production using Qwen Cloud.

Comment

Inspiration

What it does

How we built it

🎬 Character-Consistent Series Generator

Track 2: AI Showrunner — Autonomous video series generation with character consistency.

💡 Inspiration

Have you ever had a great story idea but no time or resources to turn it into an actual video? As content creators, we often struggle with:

  • Writing engaging scripts
  • Maintaining character consistency
  • Recording professional voiceovers
  • Editing and producing videos

Character-Consistent Series Generator solves all of these problems. Just type a one-line story idea, and our AI does everything else — from scriptwriting to video production — while keeping your character consistent across all scenes.

🎯 What It Does

Character-Consistent Series Generator is an AI-powered tool that turns a one-line story idea into a complete short video series with:

  • ✅ AI-Generated Script — Qwen 3.6-Plus writes a 3-scene script with descriptions, narration, and visual prompts
  • ✅ Consistent Character Design — Same character appears across all scenes using reference images
  • ✅ Professional Voiceover — Edge TTS provides natural narration in your character's voice
  • ✅ Automated Video Production — FFmpeg compiles scenes into a final video ready to watch
  • ✅ Character Reference Images — Generate or upload images from 8 different angles

🏗️ How I Built It

Architecture Overview

[User Input: One-line Story Idea] ↓ [Qwen AI: Script Generation]

3 Scenes with descriptions

Narration for each scene

Visual prompts for video ↓ [Edge TTS: Voiceover Generation]

Text-to-speech with character voice

Saves as MP3 files ↓ [FFmpeg: Video Creation]

Character image as background

Scene text overlays

Merges audio + video ↓ [Final Video Output]

All scenes combined

Ready to watch!

Technologies Used

Technology Purpose
Qwen Cloud API AI scriptwriting using qwen3.6-plus model
Edge TTS Free text-to-speech voiceover generation
FFmpeg Video processing, audio-video merging
Python 3.8+ Core programming language
DashScope Qwen model access and API integration
OpenAI Compatible API Qwen 3.6-Plus integration

🚀 Features

1. Character Consistency

  • Define your character's name, personality, style, and voice
  • Generate or upload reference images from 8 different angles
  • Same character appears in every scene

2. AI Scriptwriting

  • Input a one-line story idea
  • Qwen AI generates a 3-scene script
  • Each scene includes description, narration, and visual prompt

3. Professional Voiceover

  • Edge TTS provides natural narration
  • Multiple voice options available
  • Character voice consistency across all scenes

4. Automated Video Production

  • Creates videos for each scene
  • Merges audio and video
  • Combines all scenes into final episode

💪 Challenges I Faced

1. API Integration

Getting the Qwen Cloud API to work properly was challenging at first. I had to:

  • Understand the difference between DashScope and OpenAI-compatible endpoints
  • Correctly format API requests for script generation
  • Handle API errors and fallback gracefully

2. Video Processing

FFmpeg configuration on Windows was tricky:

  • Had to download and install FFmpeg manually
  • Debug video creation and audio merging issues
  • Ensure audio duration matched video duration

3. Character Consistency

Maintaining character consistency across scenes required:

  • Generating reference images from multiple angles
  • Using the same character image across all scenes
  • Adding text overlays for scene identification

4. File Management

Managing temporary files and cleanup:

  • Avoided deleting the final video during cleanup
  • Organized character images in reference folders
  • Handled missing files gracefully

📚 What I Learned

  1. AI Integration — Working with Qwen Cloud APIs and understanding their capabilities
  2. Video Processing — Using FFmpeg for video creation, merging, and concatenation
  3. Character Consistency — Techniques for maintaining consistent characters across scenes
  4. Error Handling — Graceful fallbacks when services fail
  5. Security — Never hardcode API keys; always use environment variables
  6. End-to-End Automation — Building a complete pipeline from idea to final product

🔮 What's Next

  • 🎥 Real Video Generation — Integrate HappyHorse or Wan2.2 for actual moving video clips
  • 🎨 Advanced Character Customization — More character parameters and styles
  • 📱 Web Interface — User-friendly web app for easier access
  • 🌍 Multi-Language Support — Scripts and voiceovers in different languages
  • 🤝 Collaboration Features — Team-based story creation and editing
  • 📈 Analytics — Track viewer engagement and content performance

📁 GitHub Repository

https://github.com/samki6576/character-series-generator

🎬 Demo Video

Watch the Demo on YouTube

🙏 Acknowledgments

  • Qwen Cloud — AI scriptwriting and image generation
  • Edge TTS — Free text-to-speech service
  • FFmpeg — Video processing framework
  • Alibaba Cloud — API infrastructure
  • Global AI Hackathon Series — Platform and community

🏆 Why This Wins

  1. 🎯 Complete Pipeline — End-to-end automation from idea to video
  2. 🎨 Character Consistency — Unique feature maintaining character across scenes
  3. 🚀 Production-Ready — Generates actual video files
  4. 💡 Creative Solution — Solves real content creation problems
  5. 🛠️ Technical Depth — Demonstrates sophisticated use of Qwen Cloud APIs

Made with ❤️ for the Global AI Hackathon Series with Qwen Cloud

Built With

  • api
  • artificial
  • automation
  • content
  • creation
  • creative
  • dashscope
  • edge
  • generation
  • generative
  • intelligence
  • llm
  • multimodal
  • natural-language-processing
  • openai
  • production
  • qwen
  • storytelling
  • text-to-speech
  • tts
Share this project:

Updates