n8n in Russia: AI models and AI agents on Claude, GPT and DeepSeek without a VPN or a foreign card
n8n in Russia can run on your own server: its code is available under the fair-code Sustainable Use License, which lets you use n8n for free for internal business and personal projects. The trouble starts with the AI nodes — they need a model provider key, and Anthropic and OpenAI don't serve Russia. With RuAPI, one OpenAI credential with the Base URL https://www.ruapi.ai/v1 is enough: the AI Agent node works with Claude, GPT, DeepSeek, Gemini and Qwen, and you top up in USDT.
$ curl https://www.ruapi.ai/v1/chat/completions \
-H "Authorization: Bearer sk-..." \
-H "Content-Type: application/json" \
-d '{
"model": "claude-sonnet-5-5",
"messages": [{"role": "user", "content": "Hello"}]
}'In n8n you set the same thing in the OpenAI credential: API Key — your RuAPI key (sk-...), Base URL — https://www.ruapi.ai/v1, with /v1 at the end. By default the Base URL field holds https://api.openai.com/v1 — replace it.
Why the n8n AI nodes are easier to connect from Russia through RuAPI
The AI nodes need a key you can't get from Russia.
Russia isn't on Anthropic's or OpenAI's supported-country lists (anthropic.com/supported-countries, developers.openai.com, Nov 1, 2026), so you can't officially set up an OpenAI or Anthropic credential with a direct key from Russia. A RuAPI key comes right after signing up with email, Google or Yandex, no phone number; you top up in USDT from 5 USDT, no foreign card needed.
A separate credential and bill for every provider.
Going direct to Claude, GPT, DeepSeek, Gemini and Qwen means several accounts, keys and payments. Through RuAPI all models run on one OpenAI credential: to switch an OpenAI Chat Model node to another model, you just type a different ID. Every request — model, tokens, charge — shows up in the console log.
Workflows run unattended — and spend unattended.
An agent triggered by a schedule or a webhook can get stuck in a loop or receive more data than expected. You can give a RuAPI key a budget: once it's used up, the key stops working even if your balance still has money. And routine steps are cheaper on light models — deepseek-flash costs $0.168 per 1M input tokens.
How to connect an AI model to n8n through RuAPI: 3 steps
Works for both self-hosted n8n and n8n Cloud: only the Base URL in the OpenAI credential changes.
Sign up and top up
Sign up at www.ruapi.ai with email, Google or Yandex. Top up in USDT or USDC from 5 USDT; USDT can be bought with rubles via SBP — guide: docs.ruapi.ai/en/buy-usdt. Foreign-issued cards are accepted too.
New accounts usually get $0.05 right away — enough to run a test workflow on claude-haiku-5-5, deepseek-flash or gpt-6-luna. Other models unlock after your first top-up.
Create a separate key for n8n
Console → API Keys → “Create API key”. For workflows that fire on their own, turn off “Unlimited” in the “Credit limit” block and enter an amount: n8n then can't spend more than that. A separate key is also easy to revoke without touching your other projects.
Add the credential and an AI Agent node
- In n8n: Credentials → New → OpenAI. API Key — your RuAPI key, Base URL — https://www.ruapi.ai/v1, with /v1 at the end (by default it holds https://api.openai.com/v1).
- Add an AI Agent node to the canvas, attach an OpenAI Chat Model to it and select the credential you created. In the model options, turn off Use Responses API: n8n works more reliably with all models (Claude, DeepSeek, Qwen, Kimi and GPT) over regular Chat Completions.
- In the Model field, pick the model from the list, or if it is not there, set the ID manually (By ID mode or an expression), exactly as on the pricing page — for example deepseek-flash or claude-sonnet-5-5.
- Run the node with a short prompt. A reply plus a new entry in RuAPI Console → Usage Logs means the setup works.
Full guide and troubleshooting: docs.ruapi.ai/en/integrations/n8n
What to enter in n8n
OpenAI credential
API Key: sk-...
Base URL: https://www.ruapi.ai/v1
Model (By ID)
deepseek-flash · qwen3.8-flash
gpt-5.4-mini · claude-haiku-5-5
claude-sonnet-5-5
Payment
USDT / USDC from 5 USDT
foreign-issued cards
no subscription
AI model prices for n8n
What models for n8n automations and agents cost — in dollars per 1M tokens, from the pricing page. No subscription: each workflow run takes only the tokens it used from your balance.
| model | input $/M | output $/M |
|---|---|---|
| deepseek-flash | $0.17 cache read $0.0120 | $1.44 |
| qwen3.8-flash | $0.14 cache read $0.0288 | $0.49 |
| 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 |
| claude-sonnet-5-5 | $2.40 cache read $0.12 | $12.00 |
| kimi-k3.1-pro | $2.40 cache read $0.24 | $12.00 |
Columns: input — new input tokens, “cache read” — input tokens served from the prompt cache, output — the model's reply. At every step the AI Agent sends the model its system message, tool descriptions and, if memory is attached, the chat history — so frequent simple tasks are cheaper on deepseek-flash, qwen3.8-flash, gpt-5.4-mini or claude-haiku-5-5. If the node sets no cache_control marks for Claude, RuAPI adds them when the request already holds earlier model replies and the previous call with the same key was less than 5 minutes ago. One-off calls and runs with long pauses are not served from cache.
The grey “>100K” line is the vendor’s rate for long prompts: if a prompt goes over the threshold, the whole request is billed at that rate.
AI in n8n: a RuAPI key or direct OpenAI and Anthropic keys
| n8n + RuAPI | Direct OpenAI / Anthropic keys | |
|---|---|---|
| Access from Russia | Works without a VPN | Russia isn't on either company's supported-country list |
| Models on one credential | Claude, GPT, DeepSeek, Gemini, Qwen, Kimi and more | Only that provider's models |
| How to connect in n8n | OpenAI credential, Base URL https://www.ruapi.ai/v1 | OpenAI credential (default https://api.openai.com/v1) or Anthropic |
Country lists: anthropic.com/supported-countries and developers.openai.com/api/docs/supported-countries; the Base URL field and its default value: the OpenAI credential code in the n8n repository (github.com/n8n-io/n8n). Checked Nov 1, 2026.