Use Cursor with any model via one gateway

Cursor can send its chat and agent requests to any OpenAI-compatible endpoint. Point the base URL override at the WebikAI gateway and your custom-model requests get routing, budgets, and receipts.

Setup

  1. Create an account and copy your wk_live_… key (shown once).
  2. Add a custom model and override the base URL:
cursor settings 
Cursor Settings → Models
1. Add a custom model named:   auto/code
2. OpenAI API Key:             wk_live_...
3. Override OpenAI Base URL:   https://gateway.webik.ai/v1
4. Click Verify

Models to start with

auto/codeCoding route ranked under your policy
auto/cheapAggressively minimize cost
auto/frontierPremium frontier route under your policy
z-ai/glm-5.2Pin an exact model (OpenRouter-style slug)

Honest limit

Cursor routes some built-in features (Tab completion, some agent modes) through its own backend regardless of this override. Only requests made with your custom model go through the WebikAI gateway — those are the ones that carry receipts and stop at your budgets.

Receipts and budgets

Every response carries five x-webik-* headers naming the request id, the model and provider that actually served it, the credential used (byok/managed), and whether a fallback fired. Requests stop at your per-key and per-client budget caps — before forwarding, never after.

Full reference: docs · live scores: benchmarks · start free