A new collaborator. Same editor.

Before you start

The current release is Orca 1.2.1 for Godot 4.7.2. Have a Godot project ready and choose a hosted model provider or a local model server.

Install Orca

1. Download the add-on.Get Orca from the Godot Asset Store, or download the latest archive from GitHub Releases.
2. Add it to your project.Ensure the add-on’s files are installed at addons/orca/ within your Godot project.
   addons/orca/Copy
3. Enable the plugin.Open your project in Godot 4.7.2. Go to Project → Project Settings → Plugins and enable Orca.
4. Open the Orca dock.Choose a provider, configure a model, and start a session.

Connect a model

Hosted providers

Open Orca’s settings, select a supported provider, and enter its credentials and a model. Credentials, endpoint, model, and reasoning-effort settings are retained separately for each provider. Provider usage fees are separate from Orca.

Local models

Start Ollama, LM Studio, or another compatible server and load a chat-capable model. Select its profile in Orca, confirm the endpoint, and refresh the model list or enter a Model ID. An API key is optional when the server does not require one.

Local and editable profiles begin in Chat mode. An isolated function-call probe must pass for the exact configuration before you can explicitly enable Agent tools.

See supported providers and local-model details →

Start with a question

In Plan mode, try a focused request such as “Explain how player movement works in this project.” Orca can inspect relevant context and help you understand the project without changing project or external state.

When you want implementation, switch to Work mode. Orca can prepare supported changes and present them for review. Read the diff and choose Apply or Reject in the editor.

Go deeper

The repository contains the full reference material for architecture, operational boundaries, and contributing.