RelayVano Pay-as-You-Go Setup: Stripe Top-Ups for Claude

Fare watch EN2026-11-01

What a RelayVano top-up actually adds to your balance

If you are meeting RelayVano for the first time, the idea is simple: you add money to a prepaid balance, and every API request draws from that balance according to the tokens it uses. There is no monthly fee, and the balance never expires, so a quiet month does not cost you anything extra. This is a pay-as-you-go arrangement, not a subscription, and you are never asked to commit to a recurring plan before you can send a request.

RelayVano is an independent relay made for developers worldwide, and that international character shows up in the interface itself. The whole site, dashboard and docs come in six languages: 中文, English, Español, हिन्दी, Português, Français. Prices show in yuan on the Chinese pages and in approximate dollars on the others, while the charge itself is made in yuan. One API key reaches every model through https://api.relayvano.com, using an OpenAI-compatible /v1 path, an Anthropic /v1/messages path, or a native Gemini /v1beta path depending on the model you call.

New accounts start with $0.40 of credit. That small starting balance is genuinely useful: it lets you run a first request, watch the numbers appear on the usage page, and confirm that your client and your chosen model are talking to RelayVano correctly before you decide how much to add.

Paying by card or WeChat Pay in Stripe checkout

Top-ups go through Stripe checkout, which means you finish the payment inside a familiar checkout flow rather than inventing a separate payment process for each region. Cards are accepted everywhere, and WeChat Pay is offered where your account supports it. If you are in China and WeChat Pay appears as an option in checkout, you can use it; if it is not offered for your account, the card route remains available.

The sequence is short. Sign in to your RelayVano dashboard, open the top-up control, choose the amount you want to add, pick a payment method, and complete the payment in Stripe checkout. Once it finishes, the funds land on your prepaid balance and are ready to use. There is no separate activation step, no coupon hunt, and no monthly renewal to cancel later.

Because the balance is prepaid and token-based, a sensible habit is to start with a modest amount while you are learning your traffic pattern, then top up again as your usage becomes predictable. Nothing is wasted while it sits there, since the balance never expires, and if you change your mind, unspent top-ups can be refunded within 9 days.

One detail worth knowing before you compare prices: the dollar figures you see on the non-Chinese pages are approximate, using a reference rate of about 1 USD ≈ ¥7.2. Your card or WeChat Pay checkout completes in yuan. Seeing the same model priced in yuan on the Chinese pages and in approximate dollars elsewhere is a convenience for readers, not two different price lists.

Model fares you can check before you top up

RelayVano charges per token, and the billing is itemised in a way that makes sense once you look at it. Prompt tokens, reply tokens, cache stores and cache hits are each priced separately, and every request shows up with its charge on the usage page, so you can trace a bill back to the call that produced it. The figures below are approximate US dollars per million tokens.

Claude Sonnet 5.5: $1.25 prompt / $6.25 reply / $0.0889 cache hit.
Claude Haiku 4.5: $0.3958 prompt / $1.90 reply / $0.0319 cache hit.
GPT-5.6 Sol: $0.3736 prompt / $2.09 reply / $0.0306 cache hit, against an official list price of $4 / $20 / $0.4 cache hit in an OpenAI promotion stated to run at least until 2026-11-21.
Gemini 3.1 Pro: $1.58 prompt / $9.13 reply / $0.1569 cache hit.
DeepSeek V4 Flash: $0.3097 prompt / $1.24 reply / $0.008333 cache hit.
MiniMax M3: $0.1472 prompt / $0.5889 reply / $0.0292 cache hit.

Here is a worked example, and it is only an example: suppose you send 200 requests a day, each one carrying 1,000 prompt tokens and receiving 500 reply tokens from MiniMax M3. The prompt portion of a single request costs $0.1472, and the reply portion adds 500 divided by a million times $0.5889, or about $0.000294. That is roughly $0.1475 per request, which works out to about $29.50 across 200 requests for the day. Swap in a larger model and the same arithmetic changes shape, which is exactly why the per-model fares are worth reading before you choose an amount to top up.

Cache hits are where careful clients can make a visible difference, and the prices above already reflect them separately. A workflow that reuses a large, stable context can pay a much smaller rate on the cached portion than on fresh prompt tokens, so it pays to check the cache hit line for the models you lean on.

One more behaviour helps when networks wobble: if a caller disconnects mid-stream, the request is still finished and billed on the tokens really used. You are not paying for tokens the model never produced, and you are not losing work that was already underway.

Which client pairs with which model

One key is convenient, but the protocol pairing is specific, and getting it right the first time saves a debugging session. 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, which means Gemini CLI, and they are not OpenAI-compatible. So a Gemini request travels over the native /v1beta path rather than the OpenAI-compatible /v1 path, while a Claude request uses the Anthropic /v1/messages path.

In practice, setup is a short conversation with your tool. Point the client at https://api.relayvano.com, paste your RelayVano API key into the credential field it already provides, and choose the model you want. Because the dashboard and docs are available in six languages, you can read the same instructions in 中文, English, Español, हिन्दी, Português or Français, which is handy when a team spans several regions.

Refunds, referrals and support once you have topped up

Try-before-you-commit is built into the terms. Unspent top-ups can be refunded within 9 days, so if you overestimated your usage in the first week, you can ask for the unused portion back rather than letting it sit indefinitely.

RelayVano also rewards people who bring colleagues along. If someone signs up through your referral link, you receive 16% of every top-up they make, and 4% of top-ups one level further down. Earnings can be withdrawn from $5.56, or moved straight into your balance and used for requests, which is a tidy option if you would rather spend the credit on tokens than on a payout.

Questions are handled from inside the product. Logged-in members write to us under Messages in the dashboard, so support history stays attached to your account and you can point at a specific request, its usage entry and its charge while you describe the problem.

Start with the $0.40 that comes with a new account, add the amount you actually need through Stripe checkout by card or by WeChat Pay where your account supports it, and point your client at the one base address. From there, the usage page tells you what each call cost, and the balance simply carries on. When you are ready, sign up for RelayVano and take your first request for a spin.