Paying for Cursor from Russia: connect your own API key

Cursor is an editor built around model use, and its subscription is paid with a foreign card. A domestic card will not do it. But the editor has a mode where it talks to an arbitrary OpenAI-compatible API using your key.

What you need

A paid Cursor Pro. Custom providers do not work on the trial — that is a limitation of the editor itself and cannot be worked around.

And a key for a compatible API. The example below uses Zukko↗, where keys look like fetch_... and top-ups go through card, bank transfer or crypto.

Setup

Open Cursor settings, go to the models section, and pick the OpenAI Compatible provider. Then two fields:

Note the /v1 suffix — it is mandatory for OpenAI-compatible mode. Without it Cursor gets a 404 on every request.

After saving, add the models you want to use. Names come from the provider's catalogue; currently enabled are claude-opus-5-5, claude-sonnet-5, gpt-5.6-sol, gpt-5.6-luna, glm-5.3 and glm-5.3-flash. The set changes; the current list is GET https://api.zukko.pro/v1/models:

curl https://api.zukko.pro/v1/models -H "Authorization: Bearer fetch_your_key"

What it costs

The arithmetic differs from a subscription. You pay per token rather than per month, so spend depends on how you use the editor.

Autocomplete and short edits cost almost nothing. What gets expensive is agent mode on a large project: it pulls many files into context, and input tokens run to hundreds of thousands per session. At $0.30 per million that is tens of cents; at Anthropic's list $4 per million (Claude Opus 5.5), a few dollars. One caveat: vendors charge ~10% of the input price for prompt-cache reads, while Zukko bills cached tokens at the regular price, so in agent mode with a large context the real gap is smaller than a straight comparison suggests.

A practical habit: keep a cheaper model on everyday tasks and reach for the heavy one deliberately. Cursor lets you add several models and switch between them in the interface.

VS Code extensions

If you do not need Cursor, the same models plug into ordinary VS Code extensions:

The rule is simple: Anthropic-compatible clients drop /v1, OpenAI-compatible ones keep it. The key is the same in every case.

If it does not work

404 on every request. Check the /v1 in the URL — Cursor needs it.

401. The key was copied with a stray space, or it has been revoked. Keys are shown once at creation; when in doubt, issue a new one.

Model not found. The name does not match the catalogue. Compare against /v1/models — case and the dots in version numbers matter.

Everything is configured but the agent will not start. You are most likely on a trial rather than Pro.

You can try it with a minimal outlay: top-ups start at $1, there is no starting balance, and subscribing to the @zukkopro↗ channel credits $0.50 once (link Telegram in the dashboard). That is enough to learn what your own working style will cost.