Are my API keys stored on the server?

In standalone mode, API keys are stored in the browser (localStorage) and never leave your device. To centralise them server-side and hide them from clients, switch to database mode with PostgreSQL — documented in our advanced deployment guide.

Are conversations saved between restarts?

In standalone mode, chat history is stored in the browser (localStorage) — it is not saved on the server. For server-side persistence, use database mode (PostgreSQL) as described in our advanced deployment guide.

Build, host and run AI solutions.

Lobe Chat

Lobe Chat — multi-model AI interface: 200+ plugins, voice synthesis, self-hosted on your VPS.

512 MB minimum (standalone mode) RAM 1 vCPU In validation

In validation

Automatic installation for this solution is ready and currently going through our tests on a real server. Ordering will open as soon as validation is complete.

Tech stack

DockerNode.jsTypeScript

Minimum RAM512 MB minimum (standalone mode)

Minimum CPU1 vCPU

Compatible OSChoice of Linux distributions

Lobe Chat is an open source AI chat interface (MIT) that brings together OpenAI, Anthropic, Google, Mistral and Ollama behind a polished design and a 200+ plugin marketplace. In standalone mode, it starts in a single Docker container with no database or mandatory authentication. Self-hosted on your ServOrbit VPS, it centralises your API keys server-side, lets you compare several providers side by side, and add custom agents with contextual memory.

Key features

Connect OpenAI, Anthropic, Google, Mistral and local Ollama from a single interface.

200+ plugin marketplace: web search, code execution, image generation, calculators and more.

Built-in voice recognition and synthesis — ask and listen hands-free.

Custom agents with system prompts, personas and contextual memory.

Side-by-side multi-model mode: compare two models in the same conversation.

Standalone Docker image — starts in seconds with no database or auth required.

When to use this solution?

1

Team AI assistant

Deploy Lobe Chat on a shared VPS to give your whole team access to multiple LLMs — API keys are centralised server-side, never exposed to browsers.

2

Model benchmarking

Compare GPT-4o, Claude Opus and a local Llama model side by side before integrating one of them into your application.

3

Developer productivity

Pair Lobe Chat with Ollama on the same VPS for a 100% local coding assistant — no external API dependency, no usage cost.

Deploy Lobe Chat on your VPS

Guide optimized for ServOrbit Cloud VPS.

01

Create the VPS

Order a ServOrbit Cloud VPS (1 vCPU / 1 GB RAM minimum). Choose Ubuntu 24.04. No domain is required to get started.

02

Install Lobe Chat

Select Lobe Chat from the ServOrbit Marketplace when ordering your VPS. Docker is pre-installed and the container starts automatically.

03

Open for the first time

Lobe Chat opens with no login required in standalone mode. Access it via an SSH tunnel (ssh -L 3210:127.0.0.1:3210 root@your-vps) then open http://localhost:3210. For a permanent HTTPS URL, attach a domain or use the free VPS subdomain from your ServOrbit dashboard.

04

Configure an LLM provider

In the interface → Settings → Providers, enter your OpenAI or Anthropic API key. To use a local model, set the Ollama URL to http://localhost:11434 if Ollama is installed on the same VPS.

05

Explore plugins and agents

Enable web search, calculators or image generators from the Plugins panel. Create a custom agent with your own system instructions and persona from the Agents section.

Frequently asked questions

No. In standalone mode Lobe Chat works without a domain via an SSH tunnel: ssh -L 3210:127.0.0.1:3210 root@your-vps then http://localhost:3210. For a permanent public URL, attach a domain — or use the free VPS subdomain from your ServOrbit dashboard — and nginx proxies Lobe Chat over HTTPS.

Add the environment variable ACCESS_CODE=your-secret when starting the container. Lobe Chat will prompt for this code on every new session. For stronger protection, place a reverse proxy with basic authentication in front of port 3210.

Yes. If Ollama is running on the same VPS, set http://localhost:11434 as the Ollama URL in Settings → Providers → Ollama. You get a fully local assistant (Llama 3, Qwen, Mistral) with no API cost. Increase VPS RAM according to the model size.

Complementary solutions

Artificial Intelligence

Lobe Chat isn't orderable yet.