Notatki inżynierskie i nowości w produkcie
Notatki zespołu, który buduje bramkę AI API Kunavo — wybór modeli, decyzje cenowe, integracja ze Stripe i SDK.
- Review·24 maja 2026·7 min czytania
Image models shootout — Nano Banana Pro vs GPT-Image-2 vs Nano Banana 2
Three of the strongest image models on Kunavo — Nano Banana Pro, GPT-Image-2, and Nano Banana 2. Best text-in-image rendering, best prompt fidelity, best value — plus real per-image cost. Pick the right one for the right use case.
- Engineering·24 maja 2026·9 min czytania
How Kunavo fails over: the four layers between an upstream error and your client
The whole failover path, with the constant behind each number: a circuit breaker that skips a failing channel before the request is sent, up to three cost-ordered hops inside one request, a stream gate that refuses to commit a 200 that turns into an error, and a four-minute backstop for an upstream that never answers.
- Guide·24 maja 2026·8 min czytania
Anthropic prompt caching — cut 90% off your input bill in 30 minutes
The full picture: how cache_control works, why the OpenAI format needs Kunavo to set the breakpoints for you, the 4-breakpoint pattern for agent loops, what silently breaks caching, and how to verify your hit rate is non-zero. Includes Kunavo cache rates per model.
- Guide·23 maja 2026·6 min czytania
Migrating from OpenAI to Kunavo in 10 minutes — Python, Node, LangChain, Vercel AI SDK
Four flavors of OpenAI integration, the one-line change each needs to start running through Kunavo, and a smoke-test that costs less than a cent.
- Guide·23 maja 2026·6 min czytania
Calling Claude with the OpenAI SDK — change one line, keep your codebase
Anthropic's SDK is great, but the ecosystem standardized on OpenAI's. Here's how to call Claude Opus 4.7, Sonnet 4.6 and Haiku 4.5 with the unmodified OpenAI Python and Node SDKs — streaming, tool use, vision included.
- Guide·23 maja 2026·7 min czytania
Veo 3 and Sora API quickstart — text-to-video and image-to-video in five minutes
Your first Veo 3 video generation calls through an OpenAI-style API: text-to-video, image-to-video with first/last frame control, file uploads, and a production-ready Python and Node example. Veo 3 is live today; Sora is on the roadmap on the same endpoint. No waitlist, no per-provider billing.
- Announcement·23 maja 2026·4 min czytania
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.