Getting Started

open-warp lets you use your own OpenAI-compatible LLM provider inside Warp through the OpenWarp.app bundle.

Install from release

Download the latest OpenWarp.app.zip from:

https://github.com/sasuke39/openwarp/releases

Unzip it, move OpenWarp.app to /Applications, then open it.

If macOS says the app is damaged, clear the quarantine attribute:

xattr -cr /Applications/OpenWarp.app

The install script also handles this automatically:

sh ./install.sh

Configure your provider

Open OpenWarp.app, then choose Settings → Agent Engine.

Fill in the fields shown in the settings window:

• provider name
• base URL
• API key
• model name

Then click save. The adapter reloads the configuration without a full app restart.

The local HTTP settings URL is still available for debugging, but normal users do not need to open it manually.

Start using AI in WarpLocal

Open OpenWarp.app, press Cmd+K, and ask a natural-language question.

Examples:

Explain the current directory.
Find the server entry point.
Create a simple Go HTTP handler and run the tests.

Chinese, Japanese, and Korean input is detected as natural language and should not fall through to the shell.