Developer integration

One familiar API surface for GPT tools

Keep the OpenAI client contract while evaluating models, routing, and cost in one account. The page is also a concise reference for English and Russian developers searching for OpenAI API alternatives.

Start with router.cheap Read the API docs Русская версия

OpenAI-compatible endpoint

Use the standard /v1/chat/completions endpoint with the official OpenAI SDK or any client that accepts a custom base URL. The same integration can be used from Python, JavaScript, cURL, Cursor, Codex, and other development tools.

from openai import OpenAI

client = OpenAI(
    api_key="sk-your-api-key",
    base_url="https://router.cheap/v1",
)
response = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[{"role": "user", "content": "Reply with OK"}],
)

A practical migration checklist

Set the base URL to https://router.cheap/v1 and keep your API key in an environment variable.
Call /v1/models to discover the models enabled for your account.
Compare output quality, latency, error handling, and total token cost on representative prompts.

Supported workflows

Use the endpoint from backend services, notebooks, CI jobs, Cursor, Codex, OpenCode, and internal tools. Keep keys server-side, set explicit timeouts, and log request IDs rather than prompts or credentials.

Start with interactive setup docs, then verify your exact model and capability requirements in the live pricing catalog.

Frequently asked questions

What base URL should I use?

Use https://router.cheap/v1 for OpenAI-compatible SDKs and the /v1/chat/completions or /v1/responses endpoint documented for your client.

Can I use this from Cursor or Codex?

Yes. Set a custom OpenAI-compatible base URL and a router.cheap API key. The integration pages below the docs explain the important settings.

Where are limits documented?

Check the live catalog, account limits, API docs, and service status before production rollout.