Integration guide

Connect LobeChat to OpenLux

LobeChat ships with an OpenAI provider that accepts a custom API host. Point that host at OpenLux and the same key unlocks every text model in the catalogue, with per-token metering.

Create an API key Read the docs

Abstract orange glass sculpture illustrating the LobeChat integration with OpenLux — LobeChat + OpenLux — one key, one endpoint, one bill.

Set it up in four steps

Create an OpenLux key

Generate a key in the console. It works for every model in the catalogue, so one key covers chat, translation and agent roles.

Set the OpenAI API host

Open Settings → Language Model → OpenAI and enter https://openlux.lol/v1 as the API proxy URL, then paste the key.

Select models per assistant

Add gpt-6-astra, claude-sonnet-4-5-20250929 and the rest as model entries — LobeChat lets you pick a different model for each assistant.

Deploy with environment variables

If you self-host LobeChat, set OPENAI_API_KEY and OPENAI_PROXY_URL so every instance starts already pointed at OpenLux.

Configuration

Copy these blocks, replace the placeholder key and you are done.

Settings → Language Model → OpenAI

API Key:      sk-openlux-...
API Proxy URL: https://openlux.lol/v1
Model List:   gpt-6-astra
              claude-sonnet-4-5-20250929
              gemini-3.1-pro-preview
              deepseek-v4

Docker (environment variables)

docker run -d -p 3210:3210 \
  -e OPENAI_API_KEY="sk-openlux-..." \
  -e OPENAI_PROXY_URL="https://openlux.lol/v1" \
  lobehub/lobe-chat

Models that fit LobeChat

GPT-6 Astra

OpenAI · $0.3677 / 1M tokens

Claude Sonnet 4.5

Anthropic · $0.2647 / 1M tokens

Gemini 3.1 Pro (preview)

Google · $0.1471 / 1M tokens

DeepSeek V4

DeepSeek · $0.0613 / 1M tokens

The full list is in the model catalogue and the use-case collections.

LobeChat + OpenLux questions

Can LobeChat use a gateway instead of the official API?

Yes. The OpenAI provider takes a custom API proxy URL, so LobeChat works with OpenLux without any plugin or fork.

How do I use several providers in one app?

Keep them all under the OpenAI provider and point the host at OpenLux — GPT, Claude, Gemini, DeepSeek and Qwen then appear as model names rather than separate integrations.

What does it cost to run LobeChat this way?

Only the tokens you use. Neither LobeChat nor OpenLux charges a seat fee, and the console shows the cost of each day and model.

Does streaming work?

Yes. OpenLux streams responses over server-sent events, which is what LobeChat expects, so responses appear token by token as usual.