Cherry Studio in Russia: Claude, GPT, Gemini and DeepSeek in one window, no VPN
Cherry Studio is a free, open-source desktop AI client for Windows, macOS and Linux: multi-model chats, assistants, translation, knowledge bases and MCP. The app itself is just the shell: answers come from whichever provider’s key you enter, and OpenAI, Anthropic and Google keys are not available from Russia. Add RuAPI as a provider in Cherry Studio and a single key gives you Claude, GPT, Gemini, DeepSeek, Qwen and GLM. Cherry Studio talks to www.ruapi.ai directly, with no VPN, 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)The cURL tab of this example is a quick way to confirm the key works before you touch the app settings. In Cherry Studio the address is shorter, https://www.ruapi.ai: the app adds /v1 itself and shows the resulting path https://www.ruapi.ai/v1/chat/completions under the field.
Why Cherry Studio does not just work in Russia
The built-in providers need a key you do not have.
OpenAI, Anthropic and Gemini are in Cherry Studio’s provider list, but each one asks for its own API key. None of the three lists Russia among its supported countries (anthropic.com/supported-countries, developers.openai.com/api/docs/supported-countries, aistudio.google.com/docs/available-regions, checked 3 Nov 2026), and they cannot be paid with a Russian card.
Every model means another account and another bill.
Cherry Studio can compare answers from several models in one conversation, but going direct means an account with every vendor. In Cherry Studio, RuAPI is a single provider: one key, a shared USDT balance, and a per-request cost record in Usage Logs in the RuAPI Console.
Through the OpenAI field Cherry Studio does not mark Claude requests for caching.
Cherry Studio only adds prompt-cache markers to Anthropic Messages requests (Cherry Studio 2.1.10 source code, 3 Nov 2026). If Claude is connected through the OpenAI field only, requests carry no markers; RuAPI then adds them, but only while you reply within 5 minutes. RuAPI has a native Anthropic address, so Cherry Studio can place the markers itself; step 4 shows how.
How to connect RuAPI to Cherry Studio: 5 steps
Menu names follow the English interface of Cherry Studio 2.1.10 (31 Oct 2026). To switch the UI language, go to Settings → Appearance → Language.
Download and install Cherry Studio
- Installers are on the official site cherryai.com/download and under Releases at github.com/CherryHQ/cherry-studio. As of 3 Nov 2026 the latest version is 2.1.10, released 31 Oct 2026.
- Windows (x64 and ARM64): the Cherry-Studio-2.1.10-win-x64-setup.exe installer, or the …-portable.exe build that runs without installing. macOS: .dmg for Apple Silicon (arm64) and Intel (x64). Linux: AppImage, .deb or .rpm for x64 and arm64.
Create a RuAPI account and key
Registering at www.ruapi.ai takes an email address or a Google or Yandex account. The balance is funded mainly in USDT, from 5 USDT on TRC20, BEP20, TON or Polygon, and USDC is fine too; if you have no crypto yet, docs.ruapi.ai/en/buy-usdt explains how to buy USDT for rubles via SBP. Cards issued by foreign banks are also accepted. A fresh account normally starts with $0.05 for claude-haiku-5-5, deepseek-flash and gpt-6-luna; the rest of the catalogue opens with your first top-up.
To get a key, open API Keys in the Console and click Create API key. There is a shortcut too: in the Console, click Connect Claude Code, Cline & more, choose Cherry Studio, then Add RuAPI to Cherry Studio, and the app sets up a RuAPI provider with your key and the OpenAI address already filled in, so step 3 can be skipped. For Claude, add the Anthropic address to that provider: on its page, click Add Endpoint next to the API Host field, in the panel that opens, click Add Endpoint again and choose Anthropic, then enter https://www.ruapi.ai and click Save.
Add RuAPI as a provider
- On first launch, click Set up another provider, then Add Provider at the bottom of the list. If Cherry Studio is already set up: Settings → Model Provider → Add Provider. The Add Custom Provider dialog opens.
- Provider Name: RuAPI. API Key: your sk-… key. Under Endpoint settings, enter https://www.ruapi.ai in the OpenAI field. "Request path: https://www.ruapi.ai/v1/chat/completions" appears under it, which means the address is accepted; if you paste it with /v1 already, the app does not add a second /v1.
- For Claude, fill in the second field as well, Anthropic, with the same https://www.ruapi.ai. The request path becomes https://www.ruapi.ai/v1/messages. Click Add.
Add models with the + button
- If Cherry Studio opens the model list by itself, click Skip. On the RuAPI provider page, click + to the right of Sync models (tooltip: Add model manually).
- In Model ID, enter the exact name, for example deepseek-flash; Model Name and Group Name fill in automatically. Click Add Model. Add the other models the same way, one at a time; their names are on the RuAPI pricing page.
- For Claude (claude-haiku-5-5, claude-sonnet-5-5 and others), set Chat protocol → Anthropic in the same dialog (this list only appears when the provider has both the OpenAI and the Anthropic field filled in). These requests go to /v1/messages, and Cherry Studio marks the system prompt and the latest messages for caching once the conversation exceeds 1024 tokens by its own estimate; each Claude model also has its own minimum cacheable length. Re-sent history is then billed at the cache read price; the first cache write costs a little more than regular input.
Check the provider and pick default models
- Click Model Check next to the API key, then Start. When the check passes, the switch at the top right turns green. If it stays grey, turn it on yourself: a disabled provider’s models do not show up in the chat.
- Settings → Default Model: choose RuAPI models for Default Assistant Model, Quick Model (Cherry Studio uses it to name chats and extract search keywords) and Translate Model. deepseek-flash or claude-haiku-5-5 is enough for the last two.
- Open Chat, pick a model from the RuAPI group at the top and send a message. The request and its cost appear right away in Usage Logs in the RuAPI Console.
Guide with screenshots: docs.ruapi.ai/en/integrations/cherry-studio
What to enter in Cherry Studio
Provider
Provider Name: RuAPI
API Key: sk-…
Endpoint settings
OpenAI: https://www.ruapi.ai
Anthropic: https://www.ruapi.ai
→ /v1/chat/completions and /v1/messages
Model (+ button)
Model ID: deepseek-flash
Claude: Chat protocol Anthropic
Payment
USDT / USDC from 5 USDT
foreign-issued cards
Model prices for Cherry Studio
A shortlist for chat, translation and Cherry Studio’s background tasks, priced in US dollars per 1M tokens exactly as on the pricing page. There is no subscription; the balance pays only for the tokens your requests use.
| model | input $/M | output $/M |
|---|---|---|
| deepseek-flash | $0.17 cache read $0.0120 | $1.44 |
| gpt-5.4-mini | $0.90 cache read $0.0900 | $5.40 |
| claude-haiku-5-5 | $0.12 cache read $0.0120 >100K: $0.60 | $0.60 >100K: $3.00 |
| gemini-3.5-flash | $1.80 cache read $0.18 | $10.80 |
| claude-sonnet-5-5 | $2.40 cache read $0.12 | $12.00 |
| gpt-5.4 | $3.00 cache read $0.30 >272K: $6.00 | $18.00 >272K: $27.00 |
| claude-opus-5-5 | $4.80 cache read $0.24 | $24.00 |
The input column is fresh prompt tokens, cache read covers the prompt portion taken from cache, output is the generated answer. Cherry Studio sends the history of the current conversation with every request, so near the end of a long chat there are far more input tokens than in the last question. GPT and DeepSeek reuse an identical request prefix from cache automatically; Claude only does so when the model is added with the Anthropic chat protocol (docs.ruapi.ai/en/prompt-caching).
The grey ">100K" row shows the vendor’s price for long contexts: once a prompt crosses the threshold, the entire request is charged at that price, not just the part above it.
Claude in Cherry Studio: Anthropic or OpenAI chat protocol
| Anthropic protocol | OpenAI protocol | |
|---|---|---|
| Provider field | Anthropic: https://www.ruapi.ai | OpenAI: https://www.ruapi.ai |
| Where the request goes | /v1/messages, native Anthropic format | /v1/chat/completions |
| Prompt caching | Cherry Studio adds cache_control markers itself | No Cherry Studio markers; RuAPI adds them if you reply within 5 minutes |
| Models | Claude only | Every RuAPI model, Claude included |
| On the trial balance | claude-haiku-5-5 | claude-haiku-5-5, deepseek-flash, gpt-6-luna |
Based on the Cherry Studio 2.1.10 source code, checked 3 Nov 2026. Both fields live in one provider, and the protocol is chosen per model, so Claude and the other models sit side by side.