AI PDF Chat Agent

An intelligent AI agent that understands and answers questions about any PDF document using Google Gemini AI and advanced retrieval techniques.

• for example i upload a resume and ask some question and you can see response below

What This Agent Does

• Reads and understands PDF document content
• Creates semantic embeddings for intelligent search
• Answers natural language questions about the PDF
• Provides accurate, context-aware responses through RAG

Prerequisites

• Node.js (v14+)
• Google Gemini API key

Quick Start

1. Install dependencies:

   npm install
   npm install @google/generative-ai @langchain/google-genai

2. Create .env file:

   GEMINI_API_KEY=your_api_key_here
   

3. Place your PDF document in data/ folder (e.g., data/document.pdf)

4. Start the AI agent:

   npx tsc
   npm run dev

The agent initializes on port 3000 and begins learning from your PDF document.

Talk to the Agent

Ask the agent questions about the PDF content:

curl -X POST http://localhost:3000/ask \
  -H "Content-Type: application/json" \
  -d '{"question": "What are the key points mentioned in this document?"}'

Agent Response

The AI agent responds with intelligent answers:

{
  "answer": "Based on the document, the key points include..."
}

Setup Agent's Knowledge

1. Visit: https://aistudio.google.com/app/apikey
2. Sign in and create a new API key
3. Add it to your .env file

Agent Architecture

src/
├── index.ts    # Agent API server
├── agent.ts    # AI agent logic
└── ingest.ts   # PDF processing
data/
└── your_document.pdf

Troubleshooting

• Check your GEMINI_API_KEY in .env
• Ensure PDF is in data/ folder
• Run npx tsc to check for compilation errors