Skip to main content

Settings & Customization

The earlier tutorials covered the core flow: workspace, model, conversation, and tools. Settings is where you manage how Helix works: where models come from, which tools the AI can use, which language the interface shows, and which defaults match your workflow.

Open Settings​

You can open settings from either place:

  • Click the gear icon in the top-right corner
  • Click Settings at the far right of the bottom toolbar

Settings page

Settings are grouped by category. When you're getting started, focus on LLM configuration, MCP tools, and Skills first.

LLM Configuration​

This is the core settings area for AI model services.

Provider Management​

Here you can:

  • View configured providers
  • Add providers such as OpenAI, Anthropic, DeepSeek, or OpenAI-compatible services
  • Edit connection details such as Base URL and API key
  • Delete providers you no longer use

Model Management​

Each provider can contain multiple models:

  • Add models: enter Model ID, display name, and context window size
  • Set a default model: choose which model new sessions use automatically
  • Tune parameters: adjust temperature, max output tokens, and related options when needed

LLM configuration

Why configure multiple providers?

You can use a cheaper model for everyday work, switch to a stronger model for complex tasks, or keep a backup provider ready when one service is unavailable.

MCP Tool Configuration​

MCP (Model Context Protocol) lets the AI connect to external tools and services. Think of it as the extension system for AI tools.

Here you can:

  • View installed MCP tools
  • Add local command-line tools or remote MCP services
  • Enable or disable individual tools
  • Edit tool parameters and launch configuration

MCP configuration

If the AI needs a capability it doesn't currently have — such as accessing an internal system, controlling a browser, or calling a custom script — MCP is usually where you extend it.

Skills Management​

Skills are reusable operating instructions and context templates that make the AI more reliable in specific scenarios.

Here you can:

  • View installed Skills
  • Install new Skills from the built-in library or custom sources
  • Create your own Skill
  • Enable or disable individual Skills

Skills management

Good Skill candidates include team project conventions, fixed release workflows, recurring troubleshooting steps, and usage notes for specific tools.

Appearance and Language​

Theme​

  • Light / dark mode: switch from the top navigation bar
  • You can also choose whether Helix follows the system theme

Language​

  • 中文 (default)
  • English

Layout​

  • Desktop / mobile layout: switch using the layout button in the top navigation bar

Advanced Options​

Once you're comfortable with the basics, you can tune these options:

SettingDescription
Thinking modeControls reasoning depth, such as Max, Standard, or Off
Output styleSwitches between concise and detailed answers
Prompt templateChanges or customizes the system prompt
Keyboard shortcutsAdjusts common keyboard operations
Keep defaults unless you know what you want to change

For most first-time users, the defaults are already good. Adjust advanced options only when you have a specific workflow issue to solve.

After the Basics​

You've now completed the core Helix onboarding path. You should know:

  • ✅ How to connect a workspace and machine
  • ✅ What each major area of the Helix interface does
  • ✅ How to configure and verify a model
  • ✅ How to start a conversation, read tool calls, and follow up
  • ✅ When to use Chat and when to use Manager
  • ✅ How to use the built-in terminal and file browser
  • ✅ Where the important settings live

Next, you can:

Happy coding with Helix! 🚀