Cursor

Cursor with GetVibe API: enable Override OpenAI Base URL in settings and paste getvibeapi.com/v1. Use gpt-6.1-sol or claude-opus-5.5.

Ready-made config

Download cursor-settings.txt

Built from our catalog on request: the API address, the recommended model and every available model. Cursor is configured in its UI only — this is a settings sheet with the address and the model IDs.

Main model — gpt-6.1-sol: strong agentic coder with a large context at a moderate price; the bare id keeps cross-channel failover on.
Small tasks — deepseek-v4.1-flash: cheapest capable model with a 1M context, for titles, summaries, commit messages and compaction.
Bare model IDs everywhere, no :suffix — that keeps automatic failover to a healthy channel on.

The catalog changes: when new models appear, download the file again. To keep only some models, add ?models= with comma-separated IDs to the URL.

Cursor is a VS Code-based IDE with built-in AI. It connects via Override OpenAI Base URL.

Setup

Open Settings → Models, enable Override OpenAI Base URL and set:

# Settings → Models → Override OpenAI Base URL
Base URL:  https://getvibeapi.com/v1
API Key:   sk-pv-…  (your key from the dashboard, pasted as is)

In the API Key field enter your key sk-pv-….

In the Settings → Models list, click + Add Custom Model. The dialog has only a NAME field: enter the public model ID character for character, then click Add (or Cancel to close it). Examples: gpt-6.1-sol, deepseek-v4.1-flash, claude-opus-5.5, claude-sonnet-5.5, google/gemini-3.8-flash, and kimi-k2.7-code.

The custom model name is the whole link: Cursor sends it in the request's model field, while the API key and Base URL come from the fields above. There is nothing else to connect. After adding it, turn on the toggle next to the model.

In the same list, turn off the built-in OpenAI models. With Override OpenAI Base URL enabled, Cursor sends everything to us, so a request for the built-in gpt-4o returns “unknown model”. That is not a key or GetVibe setup problem.

If the toggle springs back off, Cursor's own verification call failed. Send us the value of the x-getvibe-request-id header or the exact time, and we will find the request.

For normal use, write the bare model name without a :suffix: a suffix pins one channel and disables automatic switching to a healthy one.

The full model list is in the catalog.

Cursor works with /v1/chat/completions — use the base address https://getvibeapi.com/v1 (with /v1), not https://getvibeapi.com.

← PreviousOpenClaw Next →Codex Desktop