Chatbox AI in Russia: Claude, GPT, DeepSeek and Gemini on desktop and phone, no VPN
Chatbox is a free AI chat client: desktop apps for Windows, macOS and Linux, mobile apps for Android and iOS, and a web version. It has no models of its own; Chatbox simply sends your messages to whichever provider’s API key you enter. Add RuAPI as a provider in the settings and one key gives you Claude, GPT, DeepSeek, Gemini, Qwen and GLM inside Chatbox. The app talks to www.ruapi.ai directly with no VPN, no Chatbox AI subscription is needed, and the balance is funded in USDT.
Updated:
from openai import OpenAI
client = OpenAI(
api_key="sk-...",
base_url="https://www.ruapi.ai/v1",
)
resp = client.chat.completions.create(
model="deepseek-flash",
messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)This is the same request Chatbox sends on your behalf. Use the cURL tab to check your key before you touch the app settings. In Chatbox itself you only type https://www.ruapi.ai into API Host: the app adds /v1 and shows the full address https://www.ruapi.ai/v1/chat/completions in the Preview line.
What stops Chatbox from working in Russia
OpenAI, Claude and Gemini are in the provider list, but you have no key for them.
Every built-in Chatbox provider expects its own API key. OpenAI, Anthropic and Google keys cannot be obtained from Russia or paid for with a Russian card. Without a key the app opens, but no model replies.
Chatbox’s own subscription takes cards or Chinese wallets.
A Chatbox AI license is paid by card via Stripe, while WeChat Pay and Alipay are for mainland China only (chatboxai.app/en/guide/chatbox-ai/billing, 5 Nov 2026). Crypto is not on the list. The subscription also meters “compute points”, not tokens: Lite at $3.99 a month gives 2 million and excludes advanced models (chatboxai.app/pricing.md, 5 Nov 2026).
In OpenAI mode Chatbox does not mark Claude requests for caching.
Chatbox adds prompt-cache markers itself, but only in Claude API Compatible mode (Chatbox Community Edition 1.24.0 source code, 29 Oct 2026). If Claude is added as an OpenAI-compatible model, requests carry no markers; RuAPI then adds them, but only when the next message comes within 5 minutes of the previous one. Step 5 shows how to connect Claude so that Chatbox places the markers itself.
How to connect RuAPI to Chatbox: 5 steps
These steps are for the desktop app and the web version, with menu names from the English interface. To switch to Russian, go to Settings → Appearance → Language. The mobile app has a different menu layout and may not offer Russian; look for the model provider section and the same fields: API Key, API Host, Model ID.
Download Chatbox
- The official download page is chatboxai.app/en/download. Windows 10 or later (64-bit); macOS 11 or later, with separate Apple Silicon and Intel builds; Linux x86_64 and ARM64 as AppImage or .deb.
- Android: Google Play or an APK straight from the Chatbox site. iOS: the App Store. If you would rather not install anything, use the web version at web.chatboxai.app.
Sign up for RuAPI and create a key
You can open an account on www.ruapi.ai with email, Google or Yandex. New accounts usually get $0.05 to try things out, enough for claude-haiku-5-5, deepseek-flash and gpt-6-luna; the rest of the catalog unlocks after the first top-up. USDT is the easiest way to top up: from 5 USDT on TRC20, BEP20, TON or Polygon, and USDC is accepted too. USDT can be bought for rubles via SBP, step by step at docs.ruapi.ai/en/buy-usdt. Foreign-issued cards also work.
In the RuAPI Console open API Keys and click Create key. Copy the whole key; it starts with sk-.
Add a custom provider
- Open Settings → Model Provider and click Add below the provider list. In the search window pick Add Custom Provider.
- Enter RuAPI as the Name, leave API Mode set to OpenAI API Compatible and click Add.
- On the new provider’s page paste your key into API Key and type https://www.ruapi.ai into API Host. Leave API Path empty. The Preview line should read https://www.ruapi.ai/v1/chat/completions. If you paste the address with /v1 already, Chatbox will not add it twice.
Add models and test the connection
- In the Model block click New, type the exact RuAPI name into Model ID, for example deepseek-flash, and click Save. Add gpt-5.4-mini, gemini-3.8-flash and any other models the same way, one at a time.
- Fetch loads the whole RuAPI catalog at once so you can pick from it, but the list is long, so adding models by hand is usually quicker.
- Click Check and choose a model to test. Once you see Connection successful!, open a new chat, pick a model from the RuAPI provider and send a message. The cost of the request shows up straight away in the RuAPI Console under Usage Logs.
For Claude: a second provider in Claude API mode
- Click Add → Add Custom Provider again. Name it, say, RuAPI Claude and set API Mode to Claude API Compatible.
- Use the same key. In API Host enter https://www.ruapi.ai/v1, with /v1: in this mode Chatbox only appends /messages, so requests go to https://www.ruapi.ai/v1/messages, which is what Preview should show.
- Use New to add claude-haiku-5-5, claude-sonnet-5-5 or claude-opus-5-5. In this mode Chatbox marks the system prompt and the latest messages for caching on its own, so history that is sent again is usually billed at the cache-read price.
Getting started with RuAPI: docs.ruapi.ai/en
What to enter in Chatbox
Provider for all models
API Mode: OpenAI API Compatible
API Host: https://www.ruapi.ai
API Path: leave empty
Provider for Claude
API Mode: Claude API Compatible
API Host: https://www.ruapi.ai/v1
→ /v1/messages, prompt caching
Model (New button)
Model ID: deepseek-flash
Claude: claude-sonnet-5-5
Paying RuAPI
USDT / USDC from 5 USDT
foreign-issued cards
Model prices for Chatbox
Good models to start with in Chatbox, from cheap ones for everyday questions to strong ones for harder work. Prices are in US dollars per 1M tokens, as on the RuAPI pricing page. There is no monthly fee; the balance only pays for tokens.
| model | input $/M | output $/M |
|---|---|---|
| deepseek-flash | $0.17 cache read $0.0120 | $1.44 |
| claude-haiku-5-5 | $0.12 cache read $0.0120 >100K: $0.60 | $0.60 >100K: $3.00 |
| gpt-5.4-mini | $0.90 cache read $0.0900 | $5.40 |
| gemini-3.8-flash | $0.90 cache read $0.0900 | $4.50 |
| gpt-5.4 | $3.00 cache read $0.30 >272K: $6.00 | $18.00 >272K: $27.00 |
| claude-sonnet-5-5 | $2.40 cache read $0.12 | $12.00 |
| gemini-4-argon | $2.40 cache read $0.12 | $12.00 |
input is new input tokens, cache read is the part of the prompt served from cache, output is the model’s reply. Chatbox sends the current chat history with every request, so the longer the conversation, the more input tokens it uses. Start a new chat for a new topic; it costs less. GPT and DeepSeek cache automatically. For Claude, Chatbox places the markers in Claude API Compatible mode, and in OpenAI mode RuAPI adds them if no more than 5 minutes pass between messages (docs.ruapi.ai/en/prompt-caching).
The grey “>100K” row is the vendor’s long-context price: once the prompt crosses the threshold, the whole request is billed at that rate, not just the part above it.
A RuAPI key or a Chatbox AI subscription
| Your own RuAPI provider | Chatbox AI subscription | |
|---|---|---|
| Monthly fee | None, you pay per token | Lite $3.99, Pro $19.99, Pro+ $39.99 a month |
| How usage is counted | Tokens at each model’s price | Points: 2M a month on Lite, 12M on Pro |
| Which models | The full RuAPI catalog by model ID | Lite excludes advanced models |
| Payment | USDT, USDC, foreign-issued cards | Card via Stripe; WeChat Pay and Alipay for mainland China |
| Unused balance | Does not expire at month end | Point allowance is per month |
Chatbox AI plans and payment methods per chatboxai.app/pricing.md and chatboxai.app/en/guide/chatbox-ai/billing, checked 5 Nov 2026. Either way it is the same Chatbox app: the subscription is only for Chatbox’s own models, and a provider with your own key works in the free version.