I will set up a private self hosted ai server using open source models
About this gig
If you are worried about sending your business data to OpenAI, Google or other third-party AI providers, I'll set up a private AI system that runs entirely on your own hardware or server, so that you will have full control over the data.
What's included:
- Installing and configuring an open-source LLM (any model available on Hugging Face)
- Optimizing the setup for your hardware for the best possible speed
- A ChatGPT-style chat interface (Open WebUI) your team can use immediately
- Optional: self-hosting n8n and using the same models to automate tasks
I've built and benchmarked this exact setup before (30-35 tokens/sec on a 6-7 year old laptop GPU), so I can tell you whether local AI meets your needs or whether a hybrid approach makes more sense, before you spend anything.
The prices mentioned are the baseline for a single-model setup on one machine. Larger scopes (multiple models, more advanced/complicated setup, multi-user deployments) will cost more and take longer - message me with your setup and I'll send an accurate quote before you order.
Get to know Omkar Magar
AI Engineer, Building AI Automations using open source models
- FromIndia
- Member sinceAug 2026
- Avg. response time1 hour
Languages
Marathi, English, Hindi
My Portfolio
FAQ
Do I need a powerful GPU?
No - you can use budget GPU from a few years ago as well, provided it has enough VRAM to host the model of your choice. We can even offload model onto system RAM or CPU but significantly reduces the inference speeds. So I would recommend a model that fits entirely on your GPU, for faster inference
Will this replace ChatGPT entirely?
Depends the model of your choice and complexity of the task. For simple tasks like working with pdf (or other document file formats), text processing, etc. open source models are fine. But for complicated tasks like app development, large codebase changes, etc. flagship models are still far superior

