Skip to main content
NexoToken
← Back to guide center

NexoToken technical guide center

Connect Cursor to an OpenAI-compatible API

Configure an endpoint and model in Cursor, verify chat and editing, then test advanced capabilities separately.

Enter the platform's current endpoint, a dedicated credential, and the complete model ID in Cursor model settings. Verify a short blank chat before performing one reversible edit in a test project. Basic chat success does not prove completion, image, or tool support, so validate each required capability independently.

Scope

For Cursor users configuring a compatible endpoint, validating a custom model, and troubleshooting connection or editor capabilities.

Last verified

Steps

  1. Open model settings and select a custom compatible endpoint.

  2. Enter the current address, credential, and complete model ID.

  3. Send a short non-private prompt in a blank chat.

  4. Perform one reversible edit in the test project.

  5. Reconcile chat, editing, and platform usage separately.

Safe example

Base URL: https://your-api.example.com/v1
API Key: YOUR_API_KEY
Model: YOUR_MODEL_ID
Verify chat before a single-file edit.

Common errors and troubleshooting order

  1. Connection failure: inspect path and local network.
  2. Missing model: enter the complete current ID manually.
  3. Chat works but editing fails: separate client and model capabilities.
  4. Wrong record: ensure another provider was not selected.

Completion checklist

  • A short chat returns a relevant answer.
  • The test edit can be rolled back.
  • Actual model and usage match.
  • No real credential appears in project files.

Limitations and boundaries

  • Custom-endpoint and advanced-feature support varies by Cursor version.
  • A compatible protocol does not automatically provide every editor feature.