My Account
Knowledge Hub

What is Open WebUI?

The ChatGPT-killer that runs completely privately on your own infrastructure.

Open WebUI (formerly Ollama WebUI) is the premier open-source interface for interacting with Large Language Models (LLMs). It looks and feels exactly like ChatGPT, but it gives you something OpenAI doesn't: Total Sovereignty.

It connects to both local models (like Llama 3, Mistral, Gemma) and external APIs (OpenAI, Anthropic). It supports file uploads for RAG (Retrieval Augmented Generation), meaning you can upload a PDF and ask the AI questions about it without that data ever leaving your server.

⚠️ The Integration Headache

Open WebUI is just the frontend (the steering wheel). It needs an engine (Ollama) to run. If you try to host this stack yourself, you are immediately thrust into "Docker Networking Hell."

The "Docker Stack" Reality

Getting these containers to talk to each other securely is harder than it looks:

✖
Container Communication: You must configure a Docker Network so WebUI can find Ollama. If you mess up the port binding (11434), the UI loads but hangs indefinitely.
✖
Persistent Storage (RAG): If you upload documents for the AI to analyze, they are stored in a vector database inside the container. If you restart the container without mounting the correct volume, your entire knowledge base is deleted.
✖
Remote Access: WebUI is designed for localhost. If you want to access it from your phone, you have to expose it to the internet. Doing this without a proper reverse proxy and SSL certificate invites hackers to use your server for crypto mining.
✖
Inference Speed: Running Llama 3 on a standard VPS is painfully slow. You need processors with AVX-512 instruction sets (like Google C4D) to get decent token generation speeds on CPU.

The "Connection Refused" Loop

This is the most common error self-hosters face. The UI is up, but the brain is dead:

# 1. Check if Ollama is running...
root@vps:~# curl http://localhost:11434/api/tags
{"models":[{"name":"llama3:latest","size":4700000000}]}
# (It works locally...)

# 2. But the WebUI container can't see it
root@vps:~# docker logs open-webui
ERROR: Connection to http://host.docker.internal:11434 refused.
OLLAMA_BASE_URL is not reachable. Retrying in 5s...
Retrying in 5s...
Retrying in 5s...

You spend hours debugging Docker networks instead of chatting with your AI.

The Leapjuice Way

We provide a fully orchestrated AI environment. It includes the UI, the backend, and the secure storage, all running on Google Cloud's AI-Optimized hardware.

✔
Ready-to-Chat: The connection between Open WebUI and the backend is pre-configured. No port mapping required.
✔
Google C4D AI Processors: We run these instances on Google's 5th Gen AMD EPYC chips, specifically optimized for high-performance AI inference. You get fast responses without paying $500/mo for a dedicated GPU.
✔
Persistent Knowledge Base: Upload your PDFs, code snippets, and contracts. We store them on redundant Hyperdisk storage. Your "Second Brain" is safe.
✔
Secure Remote Access: Access your private AI from anywhere in the world via a secure HTTPS link protected by Cloudflare Enterprise.
Deploy Private AI

Starts at $29/mo. Private. Secure. Uncensored.

Use Cases

Private Legal/Medical Analysis: Upload sensitive documents and ask the AI to summarize them. Since the model runs on your private Leapjuice server, the data never trains a public model.

Coding Assistant: Paste proprietary code and get debugging help without leaking IP to ChatGPT.

Team Chat: Create accounts for your employees so your whole team has access to GPT-4 level intelligence for a flat monthly fee.