Browse categories
Explore
Fiverr Pro
English
$
USD
I will build a rag chatbot for your site, pdfs and database with claude code
Nisar Khan
About this gig
A chatbot that actually knows YOUR business because it reads your docs.
I build production RAG chatbots using Claude Code as the dev loop trained on your PDFs, website, database, or help center. Answers are grounded in your data, with source citations. No LLM hallucinations.
Stack: Claude Code + LangChain + FAISS / ChromaDB / Pinecone + OpenAI or Anthropic + FastAPI + your choice of frontend (web widget, Slack, WhatsApp, Telegram, Discord).
What I deliver:
- Ingestion pipeline for PDFs, Word, HTML, Notion, Confluence, SQL
- Vector DB with hybrid search (semantic plus keyword)
- Source citations on every answer
- Conversation memory plus user feedback loop
- Admin dashboard: view conversations, upload new docs, retrain
- Embed code OR Slack / WhatsApp / Telegram / Discord bot
Use cases: customer support (one shipped cut tier-1 response from 4h to 90s), internal knowledge assistant, compliance Q and A, sales enablement, technical docs search.
4 to 10 day delivery. Unlimited revisions. Message for a free scoping call.
Get to know Nisar Khan
Nisar Khan
AI Agent Developer Claude Code LangChain n8n Data Science Expert
- FromPakistan
- Member sinceDec 2022
- Avg. response time1 hour
Languages
Urdu, Pashto, English
MSc Data Science (SZABIST 2024) + IBM Certified Data Scientist.
I build production AI agents and ML systems that ship business outcomes — not demos. 20+ projects: DT International churn (92.85% ROC AUC), Amazon 4M reviews on Spark (SVM 86%), 160K real-estate forecasts with Prophet (R²=0.80), NanoLM Studio custom transformer SaaS. Stack: Claude Code, LangChain, n8n, MCP, Python, OpenAI, RAG, FastAPI, PostgreSQL, Docker. Business-level English, 1-hour response, unlimited revisions. Timezone-flexible for US, UK, EU, GCC, APAC — 7 days/week.
Free 15-min scoping call before you order.
My Portfolio
Other AI Development Services I Offer
FAQ
How is RAG different from fine-tuning?
RAG retrieves relevant chunks from your docs at query time — no retraining. You can add or update docs any time. Fine-tuning locks knowledge at training. RAG is cheaper, faster, and more controllable for knowledge bases.
Will the chatbot hallucinate answers?
Every answer includes source citations pointing to the exact passage. If a confident answer cannot be grounded, the bot says so. This is what separates RAG from pure ChatGPT.
What doc types do you ingest?
PDFs (including scanned — I use a PaddleOCR cascade from my portfolio), Word, HTML, Markdown, Notion exports, Confluence, SQL, Airtable, Google Docs.
Which LLM?
Claude, GPT-4/5, or Gemini — your choice. Claude for long-context doc QA. GPT for general. Gemini for budget.
Can you build a Slack / WhatsApp / Discord version?
Yes — all three supported natively, $120 each as extras.
Running costs after delivery?
Depends on query volume plus LLM choice. Typical small business: $20–$80/month API + $10–$30 vector DB hosting. I size LLM choice to your budget.
Can I update docs later without calling you?
Yes — Standard+ includes an admin dashboard where you upload new docs and the bot re-indexes automatically.
Biggest RAG win you have shipped?
A B2B SaaS customer-support bot with 400+ help docs: cut tier-1 ticket response from 4 hours to 90 seconds.
