Askwiseo transforms static PDFs into smart, queryable knowledge bases using powerful AI models. Built with a modern stack and optimized for speed and usability, Askwiseo is your go-to solution for turning documents into interactive, searchable experiences.
Coming soon!
- 📄 Upload and parse PDFs into structured content
- 🔍 Query the knowledge base in natural language
- 🤖 Powered by advanced LLMs (OpenAI / Hugging Face)
- 🧠 Semantic search with vector embeddings (Pinecone)
- ⚙️ Fast and scalable backend (FastAPI)
- 💾 Supabase for user management and document storage
- 🌐 Sleek and modern UI built with Next.js & Shadcn/UI
- Next.js (App Router)
- Tailwind CSS + shadcn/ui
- TypeScript
- FastAPI (Python)
- LangChain (for LLM integration and agent logic)
- Pinecone (vector database for embeddings)
- Supabase (auth + DB for metadata)
- OpenAI GPT-4 or Hugging Face Transformers (configurable)
User Uploads PDF
|
v
Text Extractor (PyMuPDF / PDFplumber)
|
v
Chunking & Embedding Generator (LangChain + OpenAI/HF)
|
v
Embeddings stored in Pinecone (vector DB)
|
v
User asks a question
-> LangChain QA chain retrieves chunks
-> LLM generates answer
-> Response shown in frontend
Bonus: Metadata is stored in Supabase (file name, upload date, user ID) for traceability.
git clone https://github.com/your-username/askwiseo.git
cd askwiseoCreate a .env file for both frontend and backend:
OPENAI_API_KEY=your_key
PINECONE_API_KEY=your_key
SUPABASE_URL=your_url
SUPABASE_SERVICE_ROLE_KEY=your_keyNEXT_PUBLIC_SUPABASE_URL=your_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_keycd backend
uvicorn main:app --reloadcd frontend
npm install
npm run devaskwiseo/
├── backend/
│ ├── main.py
│ └── services/
├── frontend/
│ ├── app/
│ └── components/
├── README.md
├── .envGot ideas to improve Askwiseo? Open a pull request or start a discussion!
MIT License.
Built with 💡 by Arun Kumar
Email: arunkumar.ai.engineer@gmail.com
Project: Askwiseo