Skip to main content

What Are Configs?

Configs are flexible containers that combine multiple components to create powerful AI coding experiences using the Continue CLI or the Extensions. A single config can include:

Models

Large Language Models for chat, autocomplete, editing, and more

MCPs

Model Context Protocol servers that provide tools and capabilities

Rules

Guidelines that shape AI behavior and responses

Prompts

Reusable instructions for common coding tasks

How Configs Work

Think of configs as recipes for AI assistance. Each config defines:
  • What models to use for different tasks (chat, autocomplete, code editing)
  • Which tools are available through MCP servers
  • How the AI behaves through rules and guidelines
  • Pre-built prompts for common workflows
This flexibility lets you create specialized setups for different contexts, like a Next.js config with React-specific rules and tools, or a data science config with Python analysis capabilities.

Config Permissions

When creating configs, you can set visibility levels:
  • Personal: Only you can see and use the config
  • Public: Anyone can discover and use your config
  • Organization: Members of your organization can access the config

Working with Configs

You can interact with configs in three main ways:
  • Create: Build a new config from scratch or start with a template
  • Edit: Modify your existing configs by adding/removing components
  • Remix: Take someone else’s config and customize it for your needs

Getting Started

1

Explore existing configs

Browse the Continue Hub to see what configs others have built
2

Add a config to your account

Click “Install Config” (+) on any config that interests you
3

Customize as needed

Add your API keys and customize components to match your workflow
4

Use in your IDE

Select the config from the dropdown in your Continue extension, type /config in the CLI, or use --config with the CLI commands

Config Format

All configs follow the config.yaml format, whether you’re using the hub interface or editing files locally. This ensures consistency between hub-managed and local configurations.
I