Cursor

Moth DocsTutorials

Cursor can send its chat to Moth through its OpenAI key setting, with some limits that come from Cursor itself.

  1. Open the key settings

    In Cursor go to Settings, then Models, then API Keys.

  2. Add the key

    Turn on OpenAI API Key and paste your moth_ key.

  3. Point it at Moth

    Turn on Override OpenAI Base URL and enter https://api.runmoth.si/v1.

  4. Add a model

    Under Models press Add model, type a Moth ID such as claude-sonnet-4.6, and turn it on.

  5. Check it worked

    Pick the model in the chat panel and ask something. The call shows in Moth Console under Logs within a few seconds.

What to expect

ChatUses your Moth key
Tab completionsKeep using Cursor's own models, since Cursor doesn't send them to your key
AgentDepends on the Cursor version, as Cursor's own forum shows, so check it on yours
Cursor's GPT modelsGo to Moth too while the override is on, so turn it off to use them again

Cursor sees your prompts. Cursor builds every prompt on its own servers before it reaches Moth, and its own retention policy doesn't cover requests made with your key. Moth keeps nothing on its side, but that promise starts where Cursor hands the request over.

Because Cursor calls from its servers, it can only reach Moth's public address, never a local one. If chat stops working after an update, turn the OpenAI API Key switch off and on and open a new chat.

A request a model can't take is refused with a 400 that says why. Some tools send stop sequences, which the GPT, Grok, DeepSeek and Kimi models refuse, so pick a Claude or Gemini model if you see that. What is removed lists every setting and what happens to it.