Integration guide

Pair-program with Aider on OpenLux

Aider reads two environment variables to reach any OpenAI-compatible provider. Point them at OpenLux and the terminal pair-programmer can use GPT-6 Astra, Claude Opus 5, DeepSeek V4 or any other model in the catalogue.

Create an API key Read the docs

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

Set it up in four steps

1

Install Aider

Install with pip or run it from a container, then start it once inside your repository.

2

Create an OpenLux key

Generate the key in the OpenLux console. Aider only needs one credential, and it covers every model in the catalogue.

3

Export the endpoint

Assign the base URL variable the OpenLux endpoint and the credential variable your OpenLux key, then launch aider with the model you want.

4

Work in the repo as usual

Aider commits changes, edits files and runs commands exactly as before — the only difference is where the tokens come from and what they cost.

Configuration

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

~/.aider.conf.yml

openai-api-base: https://openlux.lol/v1
model: openai/gpt-6-astra

Terminal

export OPENAI_API_BASE="https://openlux.lol/v1"
export OPENAI_API_KEY="sk-openlux-..."
aider --model openai/gpt-6-astra

Models that fit Aider

GPT-6 Astra

OpenAI · $0.3677 / 1M tokens

Claude Opus 5

Anthropic · $0.4412 / 1M tokens

DeepSeek V4

DeepSeek · $0.0613 / 1M tokens

GPT-5.6 Sol

OpenAI · $0.1838 / 1M tokens

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

Aider + OpenLux questions

Does Aider support custom OpenAI-compatible endpoints?

Yes. OPENAI_API_BASE overrides the endpoint for any model declared with the openai/ prefix, which is how Aider reaches OpenLux.

Why the openai/ prefix in the model name?

Aider uses the prefix to decide which client to speak with. The name after the prefix is the model ID OpenLux serves, such as gpt-6-astra or deepseek-v4.

Which model is best for large refactors?

Long refactors mean long contexts and many turns. DeepSeek V4 and GPT-5.6 Sol keep that affordable; switch to GPT-6 Astra or Claude Opus 5 for the hardest changes.

Do I need a second account for Claude or Gemini models?

No. Everything in the catalogue is reachable through the same OpenLux base URL and the same key, so Aider can switch models by changing one flag.

Connect Aider to OpenLux today

One key covers every model in this guide and the rest of the catalogue, metered in USD with free trial credit to start.

Create an API key See pricing