Open AI Micro Drama Generator
https://github.com/anil-matcha/open-ai-micro-drama-generator- Category
- AI Agents
- Rank
- No. 1342Tools index
Previous survey · No. 1324 ·
- Pricing
- Open Source
- Type
- AGENT
- Builder
- anil-matcha
- GitHub
- 486 stars
- Date
About
Turn any idea into a short-form drama via a multi-agent pipeline — screenwriter, storyboard, frames, video using Seedance, Kling, Veo, and Sora.
What it does
MicroDrama AI accepts either a concept or an existing scene script, builds characters and shots, creates reference images and video clips, then joins those clips into a downloadable MP4. A browser interface shows live progress while a Python service runs the work.
Why it's ranked here
The project makes a credible case as a runnable end-to-end prototype. It offers two input routes, live progress, structured intermediate data, and final video assembly. Its value is tempered by complete reliance on MuAPI and server state that disappears when the process restarts.
What's good
The script-first route avoids rewriting material you already have. Character portraits serve as references for later frame generation, which gives continuity an explicit mechanism. Portrait and clip work can run in parallel, while Server-Sent Events expose progress and support replay after a browser reconnects.
Tradeoffs
Every language, image, and video request depends on one external provider and its API key. Jobs live only in process memory, so restarts lose status and queued events. Generated files use local storage. The frontend proxy assumes a backend on localhost, requiring configuration work for deployment.
How to use it well
Use it for rapid story visualization, creator experiments, and turning a prepared scene into rough cinematic footage. It best suits developers comfortable running separate Python and Node services. Do not treat it as durable job infrastructure, a timeline editor, or a publishing workflow.
Technical notes+
server/api.py defines a FastAPI application with POST /api/generate, SSE at GET /api/status/{job_id}, result polling at GET /api/result/{job_id}, an in-memory jobs dictionary, permissive CORS, background pipeline execution, and static output serving. server/tools/muapi_llm.py submits requests to MuAPI, polls three seconds apart, reads MUAPI_KEY, and permits model selection through LLM_MODEL. server/utils/video.py downloads remote clips with httpx and runs MoviePy concatenation in a thread executor. client/next.config.js proxies API and output requests to localhost:8000; client/app/page.js submits generation requests and routes to a job page. client/tailwind.config.js supplies the interface theme and animations.
Observed
- License
- MIT, identified by the README badge
- Languages
- Python backend and JavaScript frontend
- Install surface
- Python requirements installation plus npm installation for the client
- Interfaces
- Next.js browser UI and FastAPI HTTP API with Server-Sent Events
- Runtime prerequisites
- Python 3.10+, Node.js 18+, and a MuAPI API key
- Output
- Locally served MP4 video assembled with MoviePy
Read from README.md, server/api.py, client/next.config.js, client/postcss.config.js, client/tailwind.config.js, client/app/page.js, client/app/layout.js, server/utils/video.py, server/tools/__init__.py, server/utils/__init__.py, server/agents/__init__.py, server/interfaces/shot.py, server/tools/muapi_llm.py, server/pipelines/__init__.py, server/interfaces/__init__.py.
What it can do
Generate screenplay from idea
Text idea or concept → Screenplay script
Create storyboard from screenplay
Screenplay script → Visual storyboard
Generate video frames from storyboard
Storyboard scenes → Individual video frames
Produce short-form drama video
Video frames and screenplay → Complete short drama video
Transform concepts into visual narratives
Abstract ideas or prompts → Structured visual story
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.