// one key, one base URL: Claude, GPT, Gemini and friends
export const flow = new OpenAI({
baseURL: "https://api.fluxorelay.com/v1",
apiKey: process.env.FLUXO_KEY,
});
export async function reviewDiff(diff: string) {
return flow.chat.completions.create({
model: "gpt-6.1-sol",
messages: [{ role: "user", content: diff }],
stream: true, // usage arrives in the last chunk
});
Tab accept the suggestion from your agent
⎇ mainTypeScriptUTF-8relay: streaming
Stay in the zone. Let the agents stream.
Fluxo relays Claude, GPT, Gemini and more through one key and one base URL, so your editor, terminal and agents never stall on billing, quotas or region locks.
$0.15 of starter credit lands in every new account.
Allowance is counted in official USD: every call draws its official price times the pass ratio.
A Claude pass covers Claude models only; a GPT pass covers GPT and Codex models only. Every other model family runs on pay-as-you-go credit.
Starting a pass creates a dedicated key for it. That key never touches your credit: once the pass is used up its calls stop with HTTP 402 until another pass is active.
A weekly cap resets 7 days after the first draw that follows the previous reset, not on calendar weeks.
Unused allowance expires with the pass and never rolls over. Each purchase is its own pass, and passes stack.
One config file is all Codex needs
Save this as ~/.codex/config.toml, export FLUXO_API_KEY with a key from your dashboard, then start codex in any repository.
.codex/config.toml
# ~/.codex/config.toml (Codex CLI and the Codex IDE extension both read it)
model = "gpt-5.6-sol" # or any other GPT id from the Fluxo price list
model_provider = "fluxo"
[model_providers.fluxo]
name = "Fluxo relay"
base_url = "https://api.fluxorelay.com/v1"
env_key = "FLUXO_API_KEY" # export FLUXO_API_KEY=flx_... before launching codex
wire_api = "responses" # Fluxo passes /v1/responses straight through