RelayVano Six-Language API Relay Setup with Claude Code in 10 Minutes

Read the RelayVano docs in your own language

Setting up an API relay usually starts with a hunt through documentation written in a language that is not your own. RelayVano removes that first hurdle. The whole site, the dashboard and the docs are available in six languages: 中文, English, Español, हिन्दी, Português, Français. Open the page in your language, and the instructions you read are the instructions we wrote for you.

That matters when you are wiring up a client for the first time. Instead of translating endpoint paths in your head, you can follow the setup steps, the client notes and the billing explanations in the language you actually think in. The documentation stays open beside your editor while you work.

Pricing follows the same idea. Prices are shown in yuan on the Chinese pages and in approximate dollars on the others, so a developer in Beijing and a developer in Buenos Aires can both compare a fare before adding funds. Nothing about reading in your own language changes what you pay for a token.

Our aim is simple: developers worldwide should be able to reach the models they need through one key, one dashboard and one usage page, without a language barrier slowing them down.

One key, three protocol routes, and the right client for each model

RelayVano is an independent relay made for developers worldwide. One API key reaches all models through https://api.relayvano.com, and the address is the same regardless of which language version of the docs you are reading. What changes is the protocol route and the client you pair with the model.

The routes are straightforward. OpenAI-compatible traffic uses /v1, Anthropic traffic uses /v1/messages, and Gemini traffic uses the native Gemini route /v1beta. You point your client at our base address, hand it your key, and choose a model.

The pairing rules are exact, and knowing them saves the most time in a ten-minute setup:

• 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. They are not OpenAI-compatible.

So if your daily driver is Claude Code, RelayVano serves it directly. If you live in Cursor or drive the terminal with cURL, you work with GPT, DeepSeek, Qwen, Kimi, MiniMax, Grok, Hunyuan and MiMo models. If you want Gemini, open Gemini CLI and keep the native protocol. There is no guessing, and no wasted afternoon.

A ten-minute RelayVano setup, from sign-up to first request

Start by creating your RelayVano account. New accounts begin with $0.40 of credit, which is enough to send a real request and see how the whole pipeline behaves before you top up. Because the dashboard is translated, this first step is the same in every supported language.

Next, open the docs in your language and copy the key from your dashboard. In your client, set the base address to https://api.relayvano.com and supply the key through the environment variable or settings field your tool expects. We keep the setup descriptions in prose rather than in code listings, so you can follow them while your terminal stays focused on your project.

Then choose the route that matches your client. Claude Code and Claude Desktop speak the Anthropic route, /v1/messages. Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs and cURL speak the OpenAI-compatible route, /v1. Gemini CLI speaks the native Gemini route, /v1beta.

Send one short request. A small prompt is enough to confirm the key, the route and the model name. When it returns, open the usage page in the dashboard. Every request appears there with its charge, so you can see exactly what the call cost before you scale it up. Ten minutes in, you have a working key, a working client and a visible bill.

Pay-as-you-go pricing you can check line by line

RelayVano sells pay-as-you-go balance. You add funds, requests are billed per token, and there is no monthly fee. The balance never expires, so a quiet week does not cost you anything and an unused top-up keeps its value until you spend it.

Prompt tokens, reply tokens, cache stores and cache hits are priced separately, and the usage page shows the charge for each request. That granularity is what makes a pay-as-you-go relay practical next to a vendor's subscription: you pay for what you actually send.

Consider an example. Say you send 200 requests a day, each with 1,000 prompt tokens and 500 reply tokens. Using GPT-6.1 Sol at $0.1958 per million prompt tokens and $0.9792 per million reply tokens, that is 200,000 prompt tokens and 100,000 reply tokens per day. The prompt side comes to about $0.0392 and the reply side to about $0.0979, roughly $0.137 a day, or about $4.11 over 30 days. This is an example of how the sums work, not a promise about your own workload.

Claude pricing is similarly transparent. Claude Fable 5.1 is listed at $4.07 prompt and $20.33 reply, against an official list price of $10 / $50, and Claude Sonnet 5.5 is listed at $1.25 prompt and $6.25 reply against an official list price of $3 / $15. Fares below are approximate US dollars per million tokens at 1 USD ≈ ¥7.2, and the charge itself is made in yuan.

One detail worth knowing early: if a caller disconnects mid-stream, we still finish the request and bill it on the tokens really used. Your cost reflects the work that was performed, not the moment your client gave up.

Top-ups, referrals and a support channel that is already open

Top-ups run through Stripe checkout. Cards work everywhere, and WeChat Pay is available where your account supports it. Unspent top-ups can be refunded within 9 days, which keeps the prepaid balance a low-risk way to begin.

If you would rather earn credit than spend it, the referral program pays 16% of every top-up made by people who sign up through your link, and 4% one level further. Earnings can be withdrawn from $5.56 or moved straight into your balance and used for requests.

Questions do not have to leave the platform. Logged-in members write to us under Messages in the dashboard, and the reply arrives in the same place you manage keys, balances and usage. If something looks off in a charge, the usage page shows the request and its price, and Messages is one click away.

Open the docs in your language and start today

RelayVano was built for developers worldwide, and the six-language site, dashboard and docs are part of that promise rather than an afterthought. Pick your language, pick your client, and let the model decide nothing about your workflow: Claude models in Claude Code or Claude Desktop, Gemini models in Gemini CLI, and GPT, DeepSeek, Qwen, Kimi, MiniMax, Grok, Hunyuan and MiMo models in the OpenAI-compatible tools you already use.

You can begin with the $0.40 of new-account credit, watch the usage page price your first request, and top up when you are ready. Sign up for RelayVano, open the docs in your own language, and have your first relay request running in about ten minutes.