LinkedIn AutoPost

An automated system that discovers AI/ML related content from the web, generates concise LinkedIn posts using a local LLM (Ollama + Qwen3:4b), and manages approval/scheduling via a Streamlit dashboard.

🚀 Features

• Scrapes AI/ML news and discussions from multiple sources
• Generates exactly 3-sentence LinkedIn posts with an educative, conversational tone
• Human-in-the-loop approval (approve/reject/edit before posting)
• Post scheduling with Celery + Redis
• Streamlit-based dashboard for easy management

🛠 Tech Stack

• Python 3.11
• FastAPI – backend services
• Celery + Redis – task queue & scheduling
• Streamlit – dashboard UI
• Ollama + Qwen3:4b – local LLM for generating posts
• Docker & Docker Compose – containerized setup

📂 Project Structure

linkedin_autopost/
│── app/                  # Core backend services
│   ├── database/         # DB connection & models
│   ├── services/         # LLM, scraping, publishing
│   ├── tasks/            # Celery tasks
│   └── utils/            # Helpers & logging
│── dashboard/            # Streamlit dashboard
│── logs/                 # Log files
│── .env                  # Environment variables (not committed)
│── docker-compose.yml    # Multi-container setup
│── requirements.txt      # Python dependencies

⚙️ Setup & Run

1. Clone the repo

   git clone https://github.com/chussboi96/linkedin_autopost.git
   cd linkedin_autopost

2. Set up environment variables
   Create a .env file (use .env.example as a reference).
   Example:

   DATABASE_URL=postgresql+psycopg2://user:password@db:5432/linkedin
   REDIS_URL=redis://redis:6379/0
   OLLAMA_BASE_URL=http://host.docker.internal:11434
   OLLAMA_MODEL=qwen3:4b

3. Run with Docker Compose

   docker compose up --build

4. Access the dashboard

   • Streamlit Dashboard → http://localhost:8501
   • API (FastAPI) → http://localhost:8000/docs

✅ Usage

• Scraper finds new content automatically.
• LLM generates draft LinkedIn posts.
• Review drafts in the dashboard → approve/edit/reject.
• Celery schedules & posts to LinkedIn.

🛡️ Security

• .env is not committed to Git (contains secrets).
• Use .env.example for reference.

📌 Future Improvements

• Multi-account support
• Richer analytics on post engagement
• Custom LLM fine-tuning

---

👤 Author

Maintained by @chussboi96