Use Relixr with other tools
OpenCode, Aider, Codex CLI, Zed, Neovim, and automation platforms like n8n — anything that accepts an OpenAI-compatible base URL.
Any tool that lets you set an OpenAI-compatible base URL and API key can run through Relixr. Prefer Connect setup packs (downloadable config + model list). The pattern is always the same: base URL is https://api.relixr.com/v1, API key is a scoped Relixr key.
OpenCode
- Open Connect → OpenCode and mint a key.
- Download
opencode.jsonand save it as~/.config/opencode/opencode.json(or a projectopencode.json). export RELIXR_API_KEY=rlx_…(key shown once in Connect).- Launch OpenCode and pick
relixr/claude-sonnet-5.
Or: npx relixr connect opencode
Aider
$ export OPENAI_API_BASE="https://api.relixr.com/v1" export OPENAI_API_KEY="rlx_your_key_here" aider --model claude-sonnet-5
Download the Aider env fragment from Connect, or run npx relixr connect aider.
Codex CLI
$ export OPENAI_BASE_URL="https://api.relixr.com/v1" export OPENAI_API_KEY="rlx_your_key_here" codex
Zed
Zed's Assistant panel accepts a custom OpenAI-compatible provider under Settings → Assistant → Add Provider. Set the base URL to https://api.relixr.com/v1 and paste your key. Use Connect → Zed for the model checklist.
Neovim (avante.nvim / CodeCompanion)
-- lua/plugins/avante.lua
{
"yetone/avante.nvim",
opts = {
provider = "openai",
openai = {
endpoint = "https://api.relixr.com/v1",
api_key_name = "RELIXR_API_KEY",
model = "claude-sonnet-5",
},
},
}Automation: n8n, LangChain, LlamaIndex
Any node or client that takes an OPENAI_API_BASE / baseURL config works the same way — set it to https://api.relixr.com/v1 and use a Relixr key as the bearer token. For workflow tools like n8n, create one key per workflow so cost attribution matches your automation, not just "n8n" as a single line item.
Automations are the highest-leverage place to set a spend cap — nobody is watching a workflow run in real time, so the cap is what stops a misconfigured loop instead of a person noticing.
Still stuck? Help center · Doctor