Skip to navigation
Coding Agents

Cursor

Cursor is an AI code editor built on VS Code. Its Override OpenAI Base URL setting sends requests from Cursor’s OpenAI slot to HyperFlowSI instead of OpenAI.

1

Create a key

Create a key for Cursor on API keys, with a spending limit.

2

Point Cursor at HyperFlowSI

  1. Open Cursor Settings (Cmd/Ctrl + ,) and go to Models.
  2. Expand API Keys and turn on OpenAI API Key.
  3. Paste your HyperFlowSI key into OpenAI API Key.
  4. Turn on Override OpenAI Base URL and set it to:
https://ai.hyperflow.fun/v1
3

Add models

Under Models, choose + Add model and enter a HyperFlowSI model id, such as openai/gpt-6-luna. Copy ids from the Models page.

4

Pick the model

Open the model picker in chat or agent and choose the model you added. Its requests now go through HyperFlowSI.

Limits

  • Tab completions always use Cursor’s own models; they don’t go through your key.
  • Cursor’s own models other than Claude and Gemini, such as Composer and Grok, are sent to HyperFlowSI too while the override is on, and fail. Turn OpenAI API Key off to use them again (Cursor forum).
  • A missing messages field error means Cursor sent a request in its newer format, which it does for some GPT models. Use a model that isn’t a GPT model, such as deepseek/deepseek-flash (Cursor forum).
  • Auto mode may not use your key. See Cursor’s docs for its current behaviour.
  • Cursor sends the requests from its own servers, so HyperFlowSI sees Cursor’s addresses, not yours.

Troubleshooting

  • Invalid API key: make sure you pasted the HyperFlowSI key (sk-...) and that it’s enabled in the console.
  • Model not found: the id must match the Models page exactly, including the lab, such as openai/gpt-6-luna.
  • Base URL: it must end in /v1.
  • 402: your balance is too low.