Start free

INTEGRATIONS · CODING AGENTS & APPS

Cursor

Cursor lets you bring your own OpenAI-compatible key. Point it at Hoonify's /v1endpoint, add the models you want, and Cursor's chat will route through Hoonify — no extension required.

Get an API key

You need a key first. Create one on the Subscriptions page — keys are revealed once at creation, so copy it before closing the dialog.
  1. Open Cursor Settings → Models (Cmd/Ctrl + Shift + J, then the Models tab).
  2. Under Models, click + Add model, type each model id below exactly (they are case-sensitive), click Add, and toggle it on. Turn off the default OpenAI models you won't use.
  3. Open API Keys, enable the OpenAI API Key toggle, and paste your Hoonify key (hoon_…).
  4. Enable Override OpenAI Base URL and set it to the Hoonify endpoint below.
Model names (case-sensitive)
deepseek-ai/DeepSeek-V4-Pro
Qwen/Qwen-3.6-27B
google/Gemma-4-31B
Override OpenAI Base URL
https://api.hoonify.ai/v1

Use the namespaced model ids

In Cursor, add models by their full namespaced, case-sensitive ids (e.g. deepseek-ai/DeepSeek-V4-Pro) — not the bare deepseek-v4-pro form used elsewhere in the API. Once added and toggled on, each appears in the model picker and works in Agent mode.

Tab stays on Cursor's models

A custom OpenAI key powers Cursor's chatand Agent on the model you add. Tab autocomplete runs on Cursor's own models and won't use your Hoonify key.

Related: OpenAI compatibility · Models