xiamuceer-j/MuMuAINovel

MuMuAINovel – AI‑powered novel‑writing assistant

What it is – A web application that lets writers create whole stories with the help of large‑language‑model APIs (OpenAI, Gemini, Claude, etc.). It provides a step‑by‑step wizard for generating outlines, characters, world‑building details and individual chapters, then lets users edit, rewrite or polish the text. All data is stored in a PostgreSQL database so multiple users can work on separate projects.

Key features

• Multi‑model support – Switch between OpenAI, Gemini, Claude or any OpenAI‑compatible proxy.
• Smart guide – AI generates plot outlines, character bios and world‑setting automatically.
• Visual management – Graphical view of character relationships, organization charts and chapter‑dependency maps.
• Chapter editor – Create, edit, re‑generate or polish sections on demand, with configurable word‑count limits.
• Role‑based login – Local username/password or LinuxDO OAuth (auto‑creates accounts).
• Production‑grade persistence – PostgreSQL with connection‑pool settings; data isolated per user.
• Docker‑first deployment – docker-compose.yml spins up PostgreSQL and the app (image mumujie/mumuainovel:latest) with a single command.
• Extensible prompts – Visual prompt‑workshop for customizing the AI’s writing style; community‑driven prompt sharing.
• Additional tools – Inspiration generator, foreshadowing tracker, “split‑book” utility, career‑level system for fantasy settings, etc.

How to get it running

1. Prerequisites – Docker & Docker‑Compose (or a Python 3.12 environment) and an API key for at least one supported LLM provider.
2. Clone & configure

   git clone https://github.com/xiamuceer-j/MuMuAINovel.git
   cd MuMuAINovel
   cp backend/.env.example .env   # edit with your DB password and LLM keys
   

3. Docker deployment (recommended)

   docker-compose up -d   # pulls the pre‑built image that already contains the embedding model
   

   The app will be reachable at http://localhost:8000.
4. If you prefer building from source – install the Python dependencies in backend/requirements.txt, start a PostgreSQL instance, run uvicorn app.main:app, then npm install && npm run dev inside frontend.

Configuration highlights

• OPENAI_API_KEY, GEMINI_API_KEY, ANTHROPIC_API_KEY – choose your provider.
• DEFAULT_MODEL defaults to gpt-4o-mini; temperature, max‑tokens, etc., are also configurable.
• Session cookies are Secure by default (set SESSION_COOKIE_SECURE=false for plain‑HTTP dev).
• Optional ALLOW_PRIVATE_AI_ENDPOINTS lets you point the app at a locally‑hosted LLM (e.g., Ollama) while keeping SSRF protection.

Hardware requirements

• Minimum: 2 CPU, 2 GB RAM, 10 GB disk (just the web service; the heavy embedding model ~400 MB is pre‑bundled in the Docker image).
• Recommended for small teams: 4 CPU, 8 GB RAM.
• For 80‑150 concurrent users: 8 CPU, 16 GB RAM, 50 GB+ SSD.

Tech stack

• Backend – FastAPI, SQLAlchemy, asyncpg, PostgreSQL, OpenAI/Claude/Gemini SDKs.
• Frontend – React 18, TypeScript, Ant Design, Zustand, Vite.
• Containerisation – Docker & Docker‑Compose.

License – GNU GPL v3 (free to use, modify and redistribute, but derivative works must remain GPL‑licensed).

Community & support

• QQ group QR code shown in the README for discussion and model‑file distribution.
• Sponsorship options (coffee, API credits) unlock perks such as a Windows‑exe launcher and priority feature requests.

---

MuMuAINovel is a genuine, open‑source AI‑assisted writing platform rather than a simple tutorial or link list.