I will build a custom rag chatbot that answers from your PDF documents with ocr


About this gig
Your best knowledge is buried in PDFs. Let's make it answer questions.
I build a complete RAG (Retrieval-Augmented Generation) assistant that reads YOUR documents and answers from them, not from guesswork. No hallucinated answers from generic ChatGPT.
WHAT YOU GET
Chat with your PDFs, with answers streamed word by word
Upload new PDFs anytime. They are searchable on the very next question, with no restart
Optional OCR for scanned PDFs (English, Bangla, Hindi)
Triple AI fallback: OpenAI Groq Ollama, so it never goes down
Conversation memory, so follow-ups like "what's the next step?" work
Answer length control: Short / Detailed / Research
Clean, professional output with no messy formatting
Modern UI in HTML, CSS and JS
TECH: Python,Html, CSS, Javascript, FastAPI, FAISS(other Vector database), OpenAI, Groq, Ollama, Tesseract(other's OCR )
PERFECT FOR
HR & policy handbooks, legal files, school syllabi, support manuals, research papers, SaaS startups.
DELIVERABLES
Full source code, .env.example, setup README (Windows/Mac/Linux).
Message me with a sample PDF before ordering and I'll tell you exactly what's possible.
Get to know MD SHAKIL HASAN
AI ENGINEER
- FromBangladesh
- Member sinceSep 2026
Languages
Bengali, English
FAQ
Which file types do you support?
PDFs, both text-based and scanned (with OCR). Other formats like DOCX or TXT can be added as an extra.
Will it make up answers?
It retrieves from your documents first, so answers stay grounded in your content.
Do I need my own API key?
Only if you want OpenAI or Groq. Ollama works free and local.
Which languages?
English, Bangla and Hindi. More can be added as an extra.
Do you provide support after delivery?
Yes. Premium includes 14 days of support, and extended support is available.

