Browse categories
Explore
Fiverr Pro
English
$
USD
Looks Like This Service Is On Hold
I will build a custom rag ai chatbot
Dumidu Sahan
About this gig
STOP using basic AI wrappers. Give your business a "Brain" that actually remembers.
I will build a professional, containerized Retrieval-Augmented Generation (RAG) system that lets you chat with your private PDFs and website data with 99.9% accuracy. Unlike standard bots, this system uses Cross-Encoder Reranking to ensure the AI never "hallucinates" on your data.
Whats Included:
- Private Data Ingestion: Upload PDFs or scrape websites instantly.
- Hybrid Intelligence: Switch between your Knowledge Base and Live Web Search.
- Enterprise Security: JWT-based Role Access (Admin vs. User).
- Persistent Memory: Sessions stay active for 24 hours via LocalStorage.
- Pro Deployment: One-click Docker Compose setup for Windows/Linux.
- High Performance: Optimized with ChromaDB and FastAPI.
️ The Tech Stack:
- Backend: Python, FastAPI, LangChain
- Frontend: Streamlit
- Vector DB: ChromaDB
- Reranker: MS-Marco Cross-Encoders
- Deployment: Docker, Railway, or VPS
Message me today for a custom demo of the system!
Get to know Dumidu Sahan
Dumidu Sahan
Professional Software Engineer
- FromSri Lanka
- Member sinceFeb 2025
- Avg. response time1 hour
Languages
Sinhala, English
Specialize in Modern Software & AI Automation
We help businesses replace messy spreadsheets with secure, high-performance systems. From Loan Management and POS to Payroll and Custom Web Apps, we build tools that scale.
NEW: Enterprise AI & RAG Development
We specialize in Retrieval-Augmented Generation (RAG), creating private AI Chatbots that chat with your PDFs/Docs with 99.9% accuracy.
🔹 Our Core: Role-based access, automated reporting, and secure workflows.
🔹 Our Edge: Cutting-edge AI integration to unlock your data.
Let's digitize and automate your business. Message us for a demo!
My Portfolio
FAQ
Will my data be used to train Public AI models (like ChatGPT)?
No. Because we use a RAG (Retrieval-Augmented Generation) architecture, your data stays in your private Vector Database. The LLM only "reads" the relevant parts to answer your specific query and does not "learn" from them permanently.
What file formats does the system support?
Out of the box, it supports PDFs. However, the system is modular and can be extended to support TXT, DOCX, and CSV files if requested.
Do I need my own API keys?
Yes. You will need a FREE Groq or OpenAI API key. This ensures you have full control over your costs and data. I will show you exactly where to get these.
Can I host this on my own company server?
Absolutely. If you choose the Premium tier, I provide a full Docker Suite. If your server supports Docker, you can launch the entire system with a single command.
Is there a limit to how many PDFs/Websites I can upload/train?
The system is built to scale. For standard business use (hundreds of documents), it is lightning-fast. For massive datasets (10,000+ pages), we can discuss a custom high-performance database setup.
