Settings and Configuration

All Chucki settings are centralized in the RAD Studio Options dialog under Tools -> Options -> Chucki.

General Settings (General)

Under Tools -> Options -> Chucki -> General you will find license management:

General settings

License

  • Sign in with your Chucki license to activate all features
  • Your license determines feature access and usage limits
  • Enterprise licenses may include additional capabilities

Detailed configurations for inline completions, models, and chat can be found in the respective sections below.

Provider Sign-Ins and Agent Chat (Chat / Models)

Under Tools -> Options -> Chucki -> Chat you will find agent-specific settings including model sign-ins:

Chat Settings

Here you can sign in to supported AI providers and configure local models for the Agent Chat:

Available Providers (Cloud)

  • Claude Code: Anthropic’s specialized code model (requires separate installation — see below)
  • GitHub Copilot: Access to Claude Sonnet, Claude Haiku, and GPT-5 models
  • OpenAI: Access to GPT-5 Codex and other OpenAI models
  • Google Gemini: Access to Google’s Gemini model via API key

Claude Code Installation

Claude Code is not bundled with the Chucki installer for licensing reasons and must be installed separately. For detailed instructions, see the Claude Code Installation Guide.

Installation options:

  • WinGet (easiest on Windows): See Claude Code guide
  • irm (Internet Resource Manager): See Claude Code guide
  • npm (Node Package Manager): Use this if you’re on Windows ARM, as WinGet and irm have issues there. See Claude Code guide

Authentication:

Claude Code currently offers no simplified authentication in the IDE. You must authenticate yourself by:

  • Direct CLI login: Run the Claude Code CLI and follow the authentication instructions (see Claude Code Guide)
  • API Key: Alternatively, you can provide an API key

Local Models for Agent Chat

Local models can also be configured here:

  1. Click Add under “Local Models”
  2. Enter the URL of your local model server (e.g., http://localhost:1234/v1)
  3. Click Check Connection — verifies connectivity and loads available models
  4. Select the desired model from the Name dropdown
  5. Save the configuration

Cloud Provider Sign-In

  1. Select your chosen provider
  2. Follow the authentication flow (browser login or API key entry)
  3. Authorize Chucki to use your account
  4. Once signed in, that provider’s models become available in the agent chat

See the Overview for details on each model’s capabilities and pricing.

Inline Completions (Completions)

Under Tools -> Options -> Chucki -> Completions you can configure Ghost Text and CodeInsight as well as local models for inline completions:

Completions Settings

  • Enable/disable inline completions
  • Choose a mode: Ghost Text or CodeInsight (mutually exclusive)
  • Local models: Configure local model connections similar to the Chat settings

If you use a local model, click Check Connection to verify availability.

MCP Server Configuration

Under Tools -> Options -> Chucki -> MCP you can configure custom MCP (Model Context Protocol) servers:

MCP settings

What Are MCP Servers?

MCP (Model Context Protocol) servers extend the Chucki Agent with additional capabilities:

  • Additional tools — Access external services and APIs
  • Context sources — Provide specialized knowledge or project data
  • Custom integrations — Connect to your internal systems

Adding MCP Servers

  1. Click the Add button in the MCP settings
  2. Enter the server details (name, endpoint, credentials if needed)
  3. Save the configuration
  4. The server’s capabilities become available in the agent chat

Use Cases

  • Connect to custom code analysis tools
  • Integrate with project management systems
  • Access company-specific documentation or databases
  • Enable specialized domain knowledge for code suggestions

Where to Next

PreviousLocal AI Models