Cursor in Russia: your own API key, no VPN, pay in rubles
Cursor works with a third-party OpenAI-compatible API: you set the address in the "Override OpenAI Base URL" field. Evrona provides exactly that — one key for GPT, Claude, Gemini, DeepSeek and the other models in the catalog.
The gateway is reachable from Russia without a VPN, and the prepaid balance is topped up in rubles with a top-up code, no foreign card required. You are charged only for the tokens you actually use.
Below: what to enter in Cursor settings, which models suit chat and agent mode, and what they cost.
Cursor lets you point at a third-party gateway: its settings carry an "Override OpenAI Base URL" field. The key works both in chat and in the agent.
What to prepare
No key yet? Create an account — keys are created in the dashboard under “API keys”, and the full value is shown only once. Copy the Model ID from the catalog in full, including the prefix before “/”.
Connection steps
Go to Settings → Models (Ctrl+, or Cmd+,).
Tick "Override OpenAI Base URL" and type the gateway address together with /v1.
Put your Evrona key into the "OpenAI API Key" field: the name is misleading, requests go to the address you entered.
Type the Model ID into the model list yourself — Cursor won't fetch the Evrona catalog.
Check it in chat first, then turn on agent mode: it rests on function calling.
Override OpenAI Base URL: https://evrona.ru/v1
OpenAI API Key: ev-YOUR_KEY
Model: google/gemini-3.7-flashTab and inline suggestions stay on Cursor's own models — that is how the editor itself is built, and your key cannot move them. The way around it for agent work is the Cline or Roo Code extension inside Cursor: it leaves the editor's own settings alone.
Function calling is supported, so agentic modes work. Strict JSON mode (response_format) and the retired functions format are turned down with an unsupported_parameter error — no parameter is lost in silence. Image input is taken by models marked "Understands images" in the catalog: a content part of type image_url with a data:image/png;base64,… URL (PNG, JPEG or WebP, up to 5 MB each, up to 50 per request). Remote URLs are not fetched, and a model without that mark returns unsupported_capability — both errors name the reason outright.
Models for Cursor
Add the Model ID from a card to the Cursor model list.
- GPT-6 Astraopenai/gpt-6-astrainput $0.67 · output $2.22 per 1M
- GPT-5.6 Solopenai/gpt-5.6-solinput $0.67 · output $3.94 per 1M
- GPT-5.6 Terraopenai/gpt-5.6-terrainput $0.29 · output $1.98 per 1M
- GPT-5.6 Lunaopenai/gpt-5.6-lunainput $0.14 · output $0.88 per 1M
- GPT-5.5openai/gpt-5.5input $0.64 · output $3.46 per 1M
- GPT-5.4openai/gpt-5.4input $0.28 · output $1.97 per 1M
Cursor FAQ
Do I need a VPN to connect Cursor to Evrona?
Evrona needs no VPN: the gateway is reachable from Russia directly. In Cursor settings, enter the gateway address with /v1 and your Evrona key.
How do I pay for Cursor models in rubles?
Top up your Evrona balance in rubles with a top-up code: buy the code in the store and redeem it in the dashboard. Each request is then charged to the balance at catalog prices.
Which models work in Cursor through Evrona?
Chat models from the Evrona catalog. Add the Model ID to the Cursor model list by hand: the editor does not fetch the Evrona catalog. For agent mode, pick a model that supports function calling.
Will Tab and inline suggestions switch to the Evrona key?
No. Tab and inline code suggestions stay on Cursor models, and your own key cannot switch them. The Evrona key works in chat and in the agent.
Can I cap what a key spends?
Yes. On the key page in the dashboard you set a spending limit per day, week or month, and a compromised key can be paused with one click.
Get a key for Cursor
Signing up takes a couple of minutes. Top up in rubles, create a key in the dashboard and paste it into the "OpenAI API Key" field.