Getting Started

What is Continue.dev?

Continue.dev is an open-source AI coding tool that plugs directly into your IDE. It works with VS Code and JetBrains IDEs, turning your code editor into an AI-assisted development environment. Unlike commercial tools, Continue.dev gives you full control over AI models and workflows through simple YAML configuration files.

How Continue.dev Works

Continue.dev uses a configuration-based system. You tell it which AI models to use, how to use them, and what commands to run. This approach offers more flexibility than tools with fixed workflows.

The extension works by intercepting your code requests and sending them to your chosen AI provider. You can use OpenAI, Anthropic, or OpenAI-compatible APIs like AskCodi. This means you can switch providers or even use multiple providers for different tasks.

Four Modes of AI Coding

Continue.dev supports four main interaction modes:

Agent Mode handles complex, multi-step tasks. Give it a high-level goal and Continue.dev will plan, execute, and refine solutions across your project.

Chat Mode provides in-context conversations about your code. Select code and ask questions. The AI sees your full context without requiring you to copy-paste.

Autocomplete offers inline code suggestions as you type. It works like enhanced IntelliSense but with full AI power behind each suggestion.

Edit Mode lets you modify specific code sections. Select code and describe what you want changed. Continue.dev will apply your edits while preserving formatting and style.

Why Choose Continue.dev Over Other Tools

GitHub Copilot and Cursor are popular but commercial. They limit you to their AI providers and workflows. Continue.dev is open source and configurable. You control which AI provider you use.

Commercial tools also lock you into their UX patterns. Continue.dev lets you customize commands, workflows, and model behavior. You can set up different configurations for different projects or teams.

The tool also supports MCP (Model Context Protocol). This connects AI models to external systems like databases, documentation, or other services. Most commercial tools don't offer this level of integration.

Installation Process

Install the Extension

Go to Extensions in VS Code (Ctrl+Shift+X). Search for "Continue". Install the official extension and reload VS Code when prompted.

For JetBrains IDEs, go to Settings > Plugins. Search for "Continue", install it, and restart your IDE.

Set Up Basic Configuration

After installation, Continue.dev creates a default configuration file. The file is usually named continue_config.yaml. The default setup includes public model configurations to help you get started.

Open the configuration file to see the structure. You'll find model definitions, context settings, and available commands. Don't worry about understanding everything at first. The defaults work fine for basic testing.

Core Configuration Explained

Continue.dev uses YAML files for configuration. A basic model configuration looks like this:

models:
  - title: "GPT-5"
    provider: "openai"
    model: "gpt-5"
    providerParams:
      apiKey: ${OPENAI_API_KEY}
    contextSizing:
      maxContextTokens: 128000

The title appears in Continue.dev's model selector. The provider tells Continue.dev which API to call. The model specifies which AI model to use. providerParams contains API authentication details.

Use environment variables like ${OPENAI_API_KEY} to store sensitive keys. This keeps secrets out of configuration files that might be shared in version control.

Essential Commands and Shortcuts

Continue.dev adds AI capabilities without changing how you work. Right-click in your editor to access the context menu. Select code and choose "Continue" commands to send selections to AI.

Continue.dev also adds slash commands. Type / in the command palette to see available commands. Common commands include:

• /ask for general questions about your code
• /edit to modify selected code
• /comment to add documentation to code
• /test to generate tests for selected functions

You can create custom slash commands in your configuration file. Define prompts that get applied to selected code. This speeds up common development tasks.

Setting Up Your First Custom Model

Configure an OpenAI-Compatible Provider

Add a new model definition to your configuration. For AskCodi API, use this structure:

models:
  - title: "AskCodi"
    provider: "openai"
    model: "askcodi-model"  # Replace with actual model name
    providerParams:
      apiKey: ${ASKCODI_API_KEY}
      baseURL: https://api.askcodi.com/v1
    contextSizing:
      maxContextTokens: 128000

Set the baseURL to AskCodi's endpoint. The tool will then route requests to AskCodi instead of OpenAI. Get your API key from the AskCodi dashboard and store it in environment variables.

Test Your Configuration

After adding the model, restart Continue.dev. The new model should appear in the model selection dropdown. Try switching between models to confirm they work.

Select some code and use Chat mode with your new model. The interaction should work the same as with any other model. If you encounter errors, check your API key and endpoint configuration.

Common Beginner Mistakes to Avoid

Starting with Continue.dev without testing configurations can lead to confusion. Always test new model configurations with simple requests before building complex workflows.

Avoid storing API keys directly in configuration files. Anyone with access to your codebase could extract the keys. Use environment variables or secure credential stores instead.

Don't try to configure everything at once. Start with basic model configurations and add features gradually. Complex configurations are harder to debug when issues arise.

Quick Takeaways

• Continue.dev integrates AI coding assistance directly in VS Code and JetBrains IDEs
• Configure AI providers through YAML files for complete control
• Use four modes: Agent, Chat, Autocomplete, and Edit based on task requirements
• Store API keys in environment variables, not configuration files
• Start with simple configurations and add complexity gradually
• Open source nature means you control your AI workflows completely
• MCP support enables integration with external systems and services

Conclusion

Continue.dev puts you in control of AI-assisted development. You choose which AI providers to use, how to configure them, and what commands to run. This level of control isn't available in commercial tools.

The configuration-based approach means you can optimize AI usage for your specific development workflow. Set up different models for different tasks. Connect to external systems through MCP. Create custom commands for your team's most common tasks.

Your next step: Install Continue.dev in VS Code and try the default configuration. Test each mode with simple requests. Then gradually add your preferred AI providers and custom configurations.