RelayVano Pay-as-you-go Setup: Claude Code and Gemini CLI

Protocols and gates EN2026-11-01

Six languages, one dashboard, and the currency on screen

RelayVano is an independent relay made for developers worldwide, and that idea shows up in the interface before you ever send a request. The whole site, the dashboard and the documentation are available in six languages: 中文, English, Español, हिन्दी, Português and Français. You can read the docs and manage your key in the language that fits your team, without translating pricing tables by hand.

The currency follows the language you choose. On the Chinese pages, prices are shown in yuan, which is also how the charge itself is settled. On the other language pages, prices appear in approximate dollars, so you can compare them against a vendor's subscription or a published list price at a glance. The underlying conversion is simple: 1 USD is roughly ¥7.2, and every figure you see on the pricing page is an approximate US dollar amount per million tokens.

That combination — six languages and a single prepaid balance — is what makes RelayVano practical for a distributed team. A colleague in Madrid and another in Hangzhou can both work from the same dashboard, in their own language, and still draw on the same balance and the same usage records.

A ten-minute setup with one API key

Getting started does not require a contract, a sales call or a monthly fee. Create your RelayVano account and you begin with $0.40 of credit, which is enough to make a first test call while you finish configuring the rest.

All traffic goes through one base address: https://api.relayvano.com. From there, three routes cover the models we offer:

An OpenAI-compatible /v1 route for GPT, DeepSeek, Qwen, Kimi, MiniMax, Grok, Hunyuan, MiMo and the other non-Claude, non-Gemini models.
An Anthropic /v1/messages route for the Claude models.
A native Gemini /v1beta route for the Gemini models, spoken by Gemini CLI.

Your workflow is short. Sign up, copy the API key from the dashboard, paste it into the environment variable your client expects, point the client at the route for your chosen model family, and send a small test request. If something looks wrong, the usage page and the documentation are both in the same six languages, so you can check the answer in the language you already chose.

Because one key reaches every model through the same address, adding a second model family later is a change of route and model name rather than a new account, a new key and a new billing relationship. That is the practical value of an API relay: one place to manage, one balance to watch.

Which client pairs with which model

Client compatibility is the part people get wrong most often, so we keep the rules short and exact.

Claude models work only in the official Claude Code and Claude Desktop clients, version 2.1.285 or newer. Claude Fable 5.1, Claude Haiku 4.5, Claude Opus 5, Claude Opus 5.5, Claude Sonnet 5 and Claude Sonnet 5.5 are reached through the /v1/messages route in those clients.
Codex, Cursor, Cherry Studio, Cline, OpenCode, the OpenAI SDKs and cURL pair only with GPT, DeepSeek and the other non-Claude, non-Gemini models, over the OpenAI-compatible /v1 route. That includes GPT-5.6 Sol, GPT-6 Astra, GPT-6.1 Sol, Qwen3.7 Plus, Qwen3.8 Max, MiniMax M3, Kimi K3, Hunyuan 4, Grok 4.6, Grok 4.7, MiMo V2.6 Pro and DeepSeek V4 Flash or V4 Pro.
Gemini models use the native Gemini protocol only, in Gemini CLI. Gemini 3.1 Pro, Gemini 3.8 Flash and Gemini 4 Argon are not OpenAI-compatible, so an OpenAI-style client is not the right place for them.

Pick the client your workflow already uses, then pick a model from the matching family. Nothing else needs to change, and no pairing outside these rules is implied.

What the token fares look like on screen

RelayVano sells prepaid balance and nothing else. You pay per token, there is no monthly fee, and the balance never expires. Prompt tokens, reply tokens, cache stores and cache hits are priced separately, and every request appears on the usage page together with its charge, so the numbers can be checked line by line.

The on-screen prices sit below many published list prices. Claude Fable 5.1 is $4.07 prompt, $20.33 reply and $0.2222 cache hit, against an official list price of $10 / $50 / $0.25 cache hit. Claude Sonnet 5.5 is $1.25 prompt, $6.25 reply and $0.0889 cache hit, against $2 / $10 / $0.1. On the GPT side, GPT-5.6 Sol is $0.3736 prompt, $2.09 reply and $0.0306 cache hit, against an official list price of $4 / $20 / $0.4 cache hit, with that OpenAI promotion stated to run at least until 2026-11-21.

For lighter traffic, MiniMax M3 is $0.1472 prompt, $0.5889 reply and $0.0292 cache hit, and Qwen3.7 Plus is $0.1458 prompt and $0.4361 reply. Gemini 3.1 Pro is $1.58 prompt and $9.13 reply, while Gemini 3.8 Flash is $1.20 prompt and $5.40 reply. All amounts are approximate US dollars per million tokens, and the charge is made in yuan.

Here is an example, not a forecast: suppose you send 200 requests a day, each with 1,000 prompt tokens and 500 reply tokens on GPT-5.6 Sol. At $0.3736 per million prompt tokens, the prompt side is about $0.0003736 per request; at $2.09 per million reply tokens, the reply side is about $0.001045. That is roughly $0.00142 per request, about $0.28 per day for 200 requests, and around $8.51 over 30 days of the same pattern. Your own mix of prompt and reply tokens will differ, which is why the usage page itemises every request.

One more billing detail worth knowing: if a caller disconnects mid-stream, the request is still finished and billed on the tokens really used. You are not charged for a response you did not receive, and you are not billed for a response that was cut off halfway either.

Top-ups, refunds, referrals and how to reach us

Adding credit goes through Stripe checkout. Cards are accepted, and WeChat Pay is available where your account supports it. If you change your mind, unspent top-ups can be refunded within 9 days.

Referrals are handled on the same prepaid model. When people sign up through your link, you receive 16% of every top-up they make, plus 4% one level further. Earnings can be withdrawn from $5.56, or moved into your balance and spent on requests.

If anything needs clarifying, logged-in members can write to us under Messages in the dashboard. Because the dashboard is available in all six languages, the conversation stays in one place and in one language.

RelayVano brings six languages, one API key and one prepaid balance together for developers worldwide. Choose your language, pick the client that matches your model family, start with the $0.40 credit already on your new account, and see the per-request charges on the usage page as you go. Sign up for RelayVano today and put your first request through https://api.relayvano.com.