DocFlow AI allows you to upload multiple PDFs, ask questions, and get AI-powered responses with document references. No more manual searches—just upload, ask, and get answers instantly!
✔ Upload multiple PDFs at once
✔ Ask questions and get answers based on document content
✔ Retrieve information from multiple files simultaneously
✔ See document names and page numbers for referenced answers
✔ Save time by avoiding manual keyword searches
I’ve spent hours searching through long PDFs—lecture notes, research papers, reports—just to find small pieces of important information.
I realized that a simple AI-powered assistant could:
✅ Help students quickly pull facts from textbooks
✅ Assist researchers in comparing studies
✅ Support professionals in analyzing legal, financial, or technical reports
That’s why I built DocFlow AI—to improve productivity, reduce manual work, and offer a smarter way to interact with documents.
📂 Upload PDFs through the chat interface.
📝 PyPDF2 extracts text from each document, even handling encrypted files.
🧩 Text is split into smaller chunks using LangChain’s Text Splitter.
🔍 Each chunk is converted into vector embeddings using Google Gemini AI and stored in a FAISS index for fast retrieval.
❓ When you ask a question, DocFlow AI searches the FAISS index to find the most relevant text.
🤖 The retrieved content is passed to Google Gemini AI, which generates a detailed response with the document name and page number.
✅ AI’s answers are displayed in the chat interface with source references for better traceability.
✅ Files Stay Available – No need to re-upload PDFs; DocFlow AI keeps them indexed for easy access.
✅ Multi-PDF Search – Query multiple PDFs at once, unlike ChatGPT or Gemini, which only work per session.
✅ Faster Retrieval – Indexed files allow instant search results without repeated reprocessing.
✅ Source Referencing – Responses clearly state the PDF name and page number for better traceability.
✅ More Control & Privacy – Files stay local or on your private server, unlike ChatGPT/Gemini, which upload them externally.
📖 Students & Researchers → Quickly extract key points from textbooks, papers, and notes.
💼 Developers & Engineers → Find technical explanations directly from manuals or API docs.
⚖️ Legal & Compliance Teams → Identify specific clauses or references across multiple legal documents.
📝 Writers & Content Creators → Summarize ideas from multiple sources.
Whether you're preparing for exams, analyzing contracts, or working with research papers, DocFlow AI makes document interaction faster and easier.
Want to try DocFlow AI yourself? Follow these steps:
- Python 3.9+
streamlitlangchainfaiss-cpugoogle-generativeaiPyPDF2python-dotenv
# Clone the repository
git clone https://github.com/devjothish/Docflow-ai.git
cd docflow-ai
# Install dependencies
pip install -r requirements.txt
# Set up API Key by creating a .env file in the root directory
echo "GOOGLE_API_KEY=your_api_key_here" > .env
# Run the app
streamlit run app.py1️⃣ Upload PDFs via the chat interface. 2️⃣ Click "Process PDFs" to create a searchable index. 3️⃣ Type a question, and DocFlow AI will retrieve relevant answers. 4️⃣ Each response shows the PDF name and page number for easy reference.
If you find DocFlow AI useful, here’s how you can contribute:
Got feedback or feature ideas? Let me know!
I built DocFlow AI because I wanted to solve a real problem—and I’m excited to share this through the AI Builders Series. Stay tuned for more AI projects!