Browse categories
Explore
Fiverr Pro
English
$
USD
I will build a secure custom rag ai knowledge base with zero API costs
M Dandi
About this gig
Stop Paying Monthly API Fees & Keep Your Company Data 100% Private!
Most AI solutions rely on expensive public APIs (like OpenAI) that expose your sensitive data and cost you money every time you ask a question.
I specialize in engineering custom RAG (Retrieval-Augmented Generation) systems using powerful open-source models via OpenRouter and fully offline deployment via Ollama (featuring models like Llama 3.2). This means you get a highly intelligent, private AI assistant with ZERO recurring API costs.
The Engineering Edge: To ensure lightning-fast retrieval speeds and eliminate AI hallucinations, my data pipeline is strictly optimized for raw .txt extraction. This technical approach drastically outperforms standard PDF processing, delivering pinpoint accuracy for your internal knowledge base.
Core Features & Technical Stack:
- 100% Secure Deployment: Built using Python and WSL for isolated, secure execution directly on your infrastructure.
- Advanced Database Integration: Combining Vector Databases with Hybrid Search (Keyword + Semantic) for enterprise-level accuracy.
- Role-Based Access Control: Strict security layers ensuring users only access the documents they are authorized to see
Get to know M Dandi
M Dandi
- FromIndonesia
- Member sinceFeb 2023
Languages
Indonesian, English
Hello everyone! Welcome to the place where creation and collaboration meet! I am M Dandi Fadli, the idea driver who is ready to help turn your vision into reality. Together, we will explore the limitless potential of every project and create stunning results. Let's start our creative adventure!
My Portfolio
FAQ
Do I need to pay for an OpenAI/ChatGPT API subscription?
No, not at all! My custom RAG system is designed to use powerful open-source models via OpenRouter or fully local models via Ollama. This means you get intelligent AI answers without any monthly API fees or token limits.
What kind of hardware do I need to run this locally?
If you want 100% offline deployment via Ollama, you will need a machine with a capable GPU and sufficient RAM, running Windows (with WSL/Ubuntu) or Linux. If your hardware is limited, don't worry—we can use OpenRouter's highly cost-effective or free cloud models instead.
Is my company's confidential data safe?
Absolutely. If we deploy the system locally using Ollama on your own hardware or server, the AI processes everything offline. Your documents never leave your internal network.
Can you build a system with multiple AI agents?
Yes! My Premium package includes setting up a multi-agent pipeline using Python. For example, we can create one agent to retrieve data, another to categorize it, and a third to track project progress automatically.
