RelayVano Pay-as-you-go API Relay for Claude Code: No Monthly Fee

1. A prepaid balance with no monthly fee and no expiry

Most of our users do not send requests at a steady, predictable pace. Maybe you ship a small tool on weekdays, run a batch job at the end of the month, or test a new prompt only when an idea strikes. That kind of irregular workload does not suit a subscription you have to keep renewing. RelayVano is built for exactly this pattern: there is no monthly fee, you pay per token, and the balance you top up never expires.

There is no clock running on your credit. If you add funds in March and only need them in August, the money is still sitting in your account when you come back. This makes RelayVano a practical fit for side projects, freelancers, students, and teams whose usage rises and falls with the work rather than with the calendar.

The billing is straightforward at the token level. Prompt tokens, reply tokens, cache stores and cache hits are each priced separately, and every request shows up on the usage page together with its charge. You can look at what you spent, on what, and when, without guessing.

One detail worth knowing: if a caller disconnects mid-stream, we still finish the request and bill it on the tokens really used. A dropped connection does not mean a lost charge or an unexplained gap in your history, so the usage page remains a reliable record of your spending.

2. One API key, three protocol routes

RelayVano is an independent relay made for developers worldwide. One API key reaches every model through https://api.relayvano.com, and the same key works across three routes: the OpenAI-compatible /v1 path, the Anthropic /v1/messages path, and the native Gemini /v1beta path.

That means you are not juggling several vendors, several invoices, or several sets of credentials. You point your client at one base address, supply one key, and choose the route that matches the model you want to call. The public endpoints are documented clearly, so setting up a new model usually comes down to changing the base address and the model name rather than rewriting your integration.

It is worth being precise about one point, because it trips people up. Gemini models are served over the native Gemini protocol only. They are not OpenAI-compatible, so a client that speaks the OpenAI format cannot simply be pointed at a Gemini model. The native route exists for exactly that purpose, and it works well with Gemini CLI.

3. Pairing each client with the models it can actually use

Client support is specific rather than approximate, and we prefer to spell it out so nothing fails halfway through your setup.

• Claude models work only in the official Claude Code and Claude Desktop clients, version 2.1.285 or newer. If you are using Claude Fable 5.1, Claude Haiku 4.5, Claude Opus 5, Claude Opus 5.5, Claude Sonnet 5 or Claude Sonnet 5.5, use one of those two clients.
• 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, including Gemini 3.1 Pro, Gemini 3.8 Flash and Gemini 4 Argon, use the native Gemini protocol only, through Gemini CLI.

There are no exceptions to those pairings, and no hidden compatibility layer that quietly changes the answer. Knowing this before you start saves the most common kind of frustration: a correctly configured key that still returns an error because the client and the model were never meant to meet.

4. What a request actually costs on the balance

Fares are shown as approximate US dollars per million tokens, with 1 USD ≈ ¥7.2, while the charge itself is made in yuan. The numbers are per million tokens, so a request that uses a few thousand tokens costs a small fraction of a listed fare.

For Claude, the comparison against official list prices is easy to read. Claude Fable 5.1 is $4.07 prompt and $20.33 reply, against an official list price of $10 / $50. Claude Haiku 4.5 is $0.3958 prompt and $1.90 reply, against an official list price of $1 / $5. Claude Sonnet 5.5 is $1.25 prompt and $6.25 reply, against $3 / $15. Cached context is priced on its own line too, with a Claude Fable 5.1 cache hit at $0.3125.

On the GPT side, GPT-6.1 Sol is $0.1958 prompt and $0.9792 reply, against an official list price of $2 / $10, while GPT-6 Astra is $0.7403 prompt and $3.70 reply, against $10 / $50. Gemini 3.8 Flash is $1.20 prompt and $5.40 reply. Among the lighter options, MiniMax M3 is $0.1472 prompt and $0.5889 reply, and Qwen3.7 Plus is $0.1458 prompt and $0.4361 reply.

Here is an example, not a forecast of your own usage. Say you send 200 requests a day, each with 1,000 prompt tokens, using MiniMax M3. That is 200,000 prompt tokens per day, or 0.2 million, so the prompt side comes to roughly $0.02944 a day before replies and cache costs. The point is not the exact figure for your project; it is that a workload like this leaves a prepaid balance almost untouched, while a heavier model such as Gemini 3.1 Pro at $1.58 prompt and $9.13 reply is still metered token by token.

5. Topping up, refunds and referrals

Top-ups run through Stripe checkout. Cards are accepted everywhere, and WeChat Pay is available where your account supports it. Because the balance never expires, a single larger top-up can cover several months of light use, and you can stop whenever you like without a renewal date to remember.

If you over-add, unspent top-ups can be refunded within 9 days. New accounts also start with $0.40 of credit, so you can try a request before deciding how much to add.

There is a referral program as well. People who sign up through your link generate 16% of every top-up for you, and 4% one level further. Earnings can be withdrawn from $5.56 or moved straight into your balance, which is a convenient way to keep working without topping up again.

6. Six languages and a dashboard you can reach

RelayVano is an independent relay made for developers worldwide, and 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, so the figures stay familiar wherever you are reading from.

Support is handled by people who know the service. If you are a logged-in member, you can write to us under Messages in the dashboard, where we can see the account and the request in question. There is no separate support portal to learn and no account number to quote from memory.

Put together, the picture is simple: one key, one base address, per-token billing with no monthly fee, and a balance that stays available until you spend it. If your usage comes in waves rather than a steady stream, that is exactly the shape RelayVano was made for. Create your account, claim the $0.40 starter credit, and send your first request through the relay.