> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tokenlab.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# Cursor

> Use TokenLab with Cursor AI code editor

## Overview

<Note>
  **Type**: Coding Tool

  **Primary Path**: OpenAI-compatible

  **Support Confidence**: Supported with limits
</Note>

[Cursor](https://cursor.sh) supports bring-your-own-key (BYOK) providers, but custom API keys only work with standard chat models. Use TokenLab here as an OpenAI-compatible endpoint for chat-style editor workflows, not as a full replacement for every Cursor model feature.

## Configuration

1. Open Cursor Settings (`Cmd/Ctrl + ,`)
2. Navigate to **Models** in the left sidebar
3. Scroll down to the **OpenAI API Key** section
4. Enter your TokenLab API key
5. Enable **Override OpenAI Base URL**
6. Set the Base URL to:

```
https://api.tokenlab.sh/v1
```

<Warning>
  Cursor custom API keys only apply to standard chat models. Tab Completion and other Cursor-managed specialized features will continue using Cursor's built-in providers.
</Warning>

<Warning>
  Cursor’s custom API key flow only applies to standard chat models. Features that require specialized Cursor-managed models, such as Tab Completion and some reasoning-specific flows, will continue using Cursor’s built-in providers. Cursor also notes that requests still pass through its backend for final prompt assembly.
</Warning>

## Model Selection

After configuration, prefer standard chat models exposed through TokenLab:

### Recommended Models for Coding

| Model        | Best For                           |
| ------------ | ---------------------------------- |
| `gpt-5.4`    | General coding, debugging, chat    |
| `gpt-5-mini` | Fast edits and lightweight prompts |

## What Works Best

### Chat and editor flows

Use TokenLab through Cursor for:

* Code explanations
* Bug fixing
* Refactoring suggestions
* Documentation generation

### What stays on Cursor-managed models

These features do not reliably switch to your TokenLab endpoint:

* Tab Completion
* Specialized non-standard chat models
* Provider-specific features that Cursor only enables on built-in models

## Troubleshooting

<AccordionGroup>
  <Accordion title="Models not showing up">
    Cursor caches model and provider state. Restart Cursor after configuring the API key and base URL.
  </Accordion>

  <Accordion title="Model unsupported in BYOK mode">
    Custom API keys only work with standard chat models. If a model uses specialized Cursor infrastructure, switch to a standard chat model such as `gpt-5.4`.
  </Accordion>

  <Accordion title="Authentication errors">
    Verify your API key is correct and has sufficient balance.
  </Accordion>

  <Accordion title="Another provider stopped working after enabling TokenLab">
    Cursor’s BYOK configuration can affect other provider flows. If another model path breaks after enabling the OpenAI base URL override, disable the override and restart Cursor.
  </Accordion>
</AccordionGroup>

## Tips

<CardGroup cols={2}>
  <Card title="Use appropriate models" icon="gauge-high">
    Use `gpt-5-mini` for quick chat-style tasks and `gpt-5.4` for heavier coding work.
  </Card>

  <Card title="Check your balance" icon="wallet">
    Monitor usage in your TokenLab dashboard.
  </Card>
</CardGroup>