Back to blog
Announcement·May 23, 2026·4 min read

Launching Kunavo — one OpenAI-compatible API for every frontier model, under official price

Why we built a Stripe-native AI API gateway: lower than upstream pricing, one bill across Claude, Gemini, GPT-Image, Veo, pay-as-you-go from a $10 top-up.

Today we're launching Kunavo — an OpenAI-compatible AI API gateway built for indie developers and small teams shipping AI features to global users.

The pitch is short: one base URL, one bearer token, frontier models below official pricing. Claude, Gemini, GPT-Image, Veo, Nano Banana, Suno — text, image, video, audio — all behind a single bill in your wallet.

Why another aggregator?

We've shipped enough AI features to know that the pain isn't the SDK. It's the operational tax around it: one contract per provider, per-model billing in different currencies, surprise rate limits on production weekends, "cheaper" aggregators that silently swap your model. Kunavo answers each of those directly.

  • Drop-in OpenAI compatibility. Streaming, function calling, tool use, vision and reasoning all work unmodified. Change one base_url.
  • Most models under the provider's official list. The margin varies per model, a few models sit exactly at list, and the per-1M-token rate for every one of them is published in plain text on /pricing. No hidden multipliers, failed requests free.
  • Stripe-native checkout. Cards everywhere Stripe reaches; Apple Pay, Google Pay and Link everywhere but India; plus local rails by country — Alipay and WeChat Pay in mainland China, Cash App Pay, Klarna and ACH in the US, KakaoPay in Korea, UPI in India, Pix in Brazil, iDEAL in the Netherlands and more on /docs/billing. $10 minimum top-up, pay-as-you-go, no subscription, balance never expires.
  • Multi-vendor hot failover. If an upstream goes down on the weekend, we re-route inside the same request — before your retry loop even fires.
  • Transparent routing. We never swap your model silently — failover switches vendor, never model. The usage dashboard breaks down every call: model, tokens, latency, status and cost.

What you get on day one

  • Claude Opus 4.7, Gemini 2.5 Pro, GPT-Image-2, Veo 3 and Nano Banana live on day one — full list at /models.
  • Native Anthropic Messages API alongside the OpenAI chat completions endpoint, so Claude users can keep their SDK.
  • Prompt caching on Claude with our affinity router keeping your cache warm — up to 90% off input tokens on repeat prompts.
  • A self-serve dashboard for keys, usage, billing, and an admin call stream that times every single request.

Three minutes to your first call

first_call.py
from openai import OpenAI

client = OpenAI(
    api_key="sk-kn-...",
    base_url="https://api.kunavo.com/v1",  # change only this line
)

# Same SDK. Different model.
resp = client.chat.completions.create(
    model="claude-sonnet-4-6",
    messages=[{"role": "user", "content": "Hello, Kunavo"}],
)
print(resp.choices[0].message.content)

Sign up at kunavo.com/app/signup — top up from $10, pay-as-you-go, enough to put every modality through its paces.

What's next

On the roadmap for the next two months: an in-product migration assistant from OpenAI, Anthropic and OpenRouter; embeddings-aware token batching; and team accounts with per-key budgets and per-environment audit logs. We'll post each ship in /changelog.

Bugs and product feedback land in our inbox at contact@kunavo.com. We answer every email.

Related deep dives