RelayVano Claude Code API Relay Pricing: 9-Day Refund Window

Fare watch EN2026-11-01

What the 9-day refund window actually covers

Adding money to a developer API account is easier when you know what happens if your plans change. At RelayVano, top-ups are prepaid, you pay per token, there is no monthly fee, and the balance never expires. If you top up and leave money unspent, you can request a refund within 9 days.

The useful way to read that policy is without small print: the 9-day window applies to unspent top-ups. It gives you time to try the service, check the usage page, and decide whether the pricing and model mix suit your work before the money is tied up. Once you have used part of a top-up, that request is billed on the tokens really used.

That last point matters for your decision. RelayVano bills prompt tokens, reply tokens, cache stores, and cache hits separately, and every request appears with its charge on the usage page. You can therefore review actual spending rather than estimate it from a vague subscription. If a request disconnects mid-stream, we still finish the request and bill it on the tokens really used.

Here is the first question for your checklist: would you rather control spending request by request? With RelayVano, you do not need to commit to a vendor's subscription. You top up the amount you expect to need, watch the balance, and use the 9-day refund window while part of your top-up is still unspent.

Claude Code, Gemini CLI, and the client you already use

The next decision is not about price alone. It is about whether the model you want can speak the protocol your client expects. RelayVano is an independent relay made for developers worldwide, with one API key reaching every model through https://api.relayvano.com.

Claude models work only in the official Claude Code and Claude Desktop clients, version 2.1.285 or newer. Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs, and cURL pair only with GPT, DeepSeek, and the other non-Claude, non-Gemini models. Gemini models use the native Gemini protocol only, in Gemini CLI; they are not OpenAI-compatible.

That sounds strict because it is specific. Before you sign up, list the clients you actually use and the models you actually want to call:

If your workflow runs in Claude Code or Claude Desktop, Claude models are the relevant choice.
If you work in Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs, or cURL, use GPT, DeepSeek, and the other supported non-Claude, non-Gemini models.
If your workflow is built around Gemini CLI, use Gemini models through the native Gemini protocol.

RelayVano also provides an OpenAI-compatible /v1 route, an Anthropic /v1/messages route, and native Gemini /v1beta access. The point is not that one client can somehow run every protocol. The point is that one RelayVano key can cover your projects, provided each model is paired with its supported client.

Check the token fares against your own workload

Pay-as-you-go pricing is easiest to judge when you look at the same three numbers for the models on your shortlist. RelayVano shows approximate US dollars per million tokens, with 1 USD ≈ ¥7.2, while the actual charge is made in yuan.

For Claude, the fares below make the value visible. Claude Sonnet 5.5 is $1.25 prompt, $6.25 reply, and $0.0889 cache hit, compared with an official list price of $2 / $10 / $0.1 cache hit. Claude Opus 5.5 is $1.65 prompt, $7.82 reply, and $0.075 cache hit, compared with $4 / $20 / $0.2 cache hit. Claude Haiku 4.5 is $0.3958 prompt, $1.90 reply, and $0.0319 cache hit, compared with $1 / $5 / $0.1 cache hit.

You can also compare a GPT route with Claude. GPT-6.1 Sol is $0.1958 prompt, $0.9792 reply, and $0.0389 cache hit, against an official list price of $2 / $10 / $0.1 cache hit. GPT-5.6 Sol is $0.3736 prompt, $2.09 reply, and $0.0306 cache hit, against $4 / $20 / $0.4 cache hit; that official price is described as an OpenAI promotion stated to run at least until 2026-11-21.

For a concrete example, suppose you send 200 requests a day and each request uses 1,000 prompt tokens and 500 reply tokens. With Claude Sonnet 5.5, the daily prompt cost would be 200,000 tokens, or $0.25, and the daily reply cost would be 250,000 tokens, or $1.5625. The example totals $1.8125 before cache charges, or about $50.69 across 28 days. Your real cost depends on the tokens your prompts and replies actually contain, and cache stores and cache hits are priced separately.

That example is a calculation, not a promise. It shows why the usage page is a better budgeting tool than a monthly estimate. You can check each request and its charge, then adjust your top-up accordingly.

Languages, payment methods, and everyday support

RelayVano was made for developers worldwide. The whole site, dashboard, and documentation come in six languages: 中文, English, Español, हिन्दी, Português, Français. Prices appear in yuan on the Chinese pages and in approximate dollars on the others. If your team reads or writes in one of those languages, you can navigate the product and review charges without translating a technical interface by hand.

Top-ups go through Stripe checkout. Cards work everywhere, and WeChat Pay is available where your account supports it. New accounts start with $0.40 of credit, which is enough to make a first request and inspect the usage page before you add more.

Support is also designed to stay inside the product. Logged-in members write to us under Messages in the dashboard. If a model pairing is unclear, a charge needs context, or you want to confirm how a top-up was recorded, you can ask there rather than searching for a separate support channel.

Your checklist should include four practical items: the language your team wants, the payment method your account supports, whether $0.40 is enough for an initial test, and where you will send questions. RelayVano keeps these steps in the dashboard, so the evaluation can happen in the same place where you will monitor usage.

Use referrals carefully, then make your decision

RelayVano also offers a referral programme. You receive 16% of every top-up by people who sign up through your link, plus 4% one level further. Earnings can be withdrawn from $5.56 or moved into the balance. If you already use RelayVano, the referral link gives you a way to offset some of your own top-up cost while helping a colleague try the service.

Before adding funds, run through the full decision once more:

Confirm that the model you want matches your client: Claude with Claude Code or Claude Desktop, Gemini with Gemini CLI through its native protocol, or a supported non-Claude, non-Gemini model with Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs, or cURL.
Compare prompt, reply, and cache-hit fares for the models you expect to use.
Remember that there is no monthly fee and the prepaid balance never expires.
Use the 9-day refund window for unspent top-ups if your test period changes your mind.
Review the usage page, where every request and its charge are shown.

That is the complete picture: a worldwide, six-language relay, one API key across the supported routes, transparent per-token pricing, and a 9-day refund window on unspent top-ups. Start with the $0.40 new-account credit, test the model and client pairing you need, and watch the first charges appear. If RelayVano fits your workflow, sign up and top up with the amount you are comfortable trying.