An end-to-end LLM-powered research & writing companion that turns any topic into a polished, SEO-ready blog post. It fetches real world data using tavily to use upto date information for the blog. The assistant mimics how a human content writer would plan, research and execulte a blog post. Each node is assigned a defined role in the workflow to get a coherent blog.
- 🧠 Smart Topic Analysis – automatically decomposes complex topics.
- 🔍 Real-time Web Research – fetches up-to-date sources via Tavily
- ✍️ Structured Drafting – consistent tone and real world info with minimum hallucinations.
- 🎯 Interactive Chat UI – Thanks to the Chainlit for providing a nice UI.
- 💬 Memory-Backed ChitChat – context-aware conversations with
mem0. It is basically a RAG over past conversations and provides the ChitChat node with context aware past interactions
| Node | Purpose |
|---|---|
| Intent Classification | Detects whether you want to create, edit, or just chat |
| Topic Analysis | Extracts the main subject and sub-topics |
| Question Generation | Produces research questions to cover |
| Web Search | Retrieves answers & sources |
| Content Writing | Builds the first draft |
| Content Editing | Refines or rewrites based on your feedback |
- Python ≥ 3.12
- API keys: Groq (Free for limited usage)
- Tavily (Free for limited usage)
- Mem0 (Free for limited usage)
- LangSmith (optional)
# 1. Clone
git clone https://github.com/ahsannawazch/ai-blog-assistant
cd ai-blog-assistant
# 2. Install
pip install poetry
poetry install
# 3. Configure
cp .env.sample .env
# edit .env with your keys
# 4. Run
chainlit run app.py --port 8000
# open http://localhost:8000Found a bug or have an idea?