Kodely Docs

Features

Kodely comes with a rich set of features designed to enhance your coding experience. This section covers the key features that make Kodely a powerful AI coding assistant.

API Configuration Profiles

API Configuration Profiles allow you to set up and switch between different AI providers and models.

Creating Profiles

To create a new API configuration profile:

  1. Open the Kodely panel
  2. Click on the Settings icon (gear)
  3. Select "API Configuration"
  4. Click "Add New Profile"
  5. Enter a name for the profile
  6. Select the provider (Anthropic, OpenAI, etc.)
  7. Enter your API key
  8. Configure any additional settings
  9. Click "Save"

Switching Between Profiles

To switch between profiles:

  1. Click on the model selector dropdown in the Kodely panel
  2. Select the profile you want to use
  3. Kodely will immediately switch to using that profile

Use Cases

  • Multiple providers: Use different AI providers for different tasks
  • Different models: Switch between faster models for simple tasks and more powerful models for complex tasks
  • Team sharing: Set up profiles for different team members or projects
  • Fallback options: Have backup providers in case one is unavailable

Auto-Approving Actions

Auto-approval settings allow you to streamline your workflow by automatically approving certain actions without requiring explicit confirmation each time.

Configuring Auto-Approval

To configure auto-approval settings:

  1. Open the Kodely panel
  2. Click on the Settings icon (gear)
  3. Select "Auto-Approval Settings"
  4. Toggle the switches for the actions you want to auto-approve
  5. Click "Save"

Available Auto-Approval Options

  • Read files: Automatically allow Kodely to read files in your workspace
  • Create files: Automatically allow Kodely to create new files
  • Modify files: Automatically allow Kodely to modify existing files
  • Run safe commands: Automatically allow Kodely to run commands that are considered safe
  • Web searches: Automatically allow Kodely to perform web searches

Security Considerations

When using auto-approval, keep these security considerations in mind:

  • Only enable auto-approval for actions you're comfortable with
  • Consider the sensitivity of your codebase before enabling file modifications
  • Potentially destructive commands will still require explicit approval
  • You can always disable auto-approval if you need more control

Custom Instructions

Custom instructions allow you to personalize Kodely's behavior to better match your preferences and needs.

Setting Custom Instructions

To set custom instructions:

  1. Open the Kodely panel
  2. Click on the Settings icon (gear)
  3. Select "Custom Instructions"
  4. Enter your instructions in the text area
  5. Click "Save"

Effective Custom Instructions

Here are some examples of effective custom instructions:

  • Coding style: "Always use functional components in React and prefer arrow functions"
  • Documentation: "Include detailed comments for complex functions and follow JSDoc format"
  • Testing: "Write unit tests for all new functions using Jest"
  • Project specifics: "This is a Next.js project using TypeScript and Tailwind CSS"
  • Personal preferences: "I prefer concise explanations with code examples"

Global vs. Mode-Specific Instructions

You can set both global instructions that apply to all modes and mode-specific instructions:

  • Global instructions: Apply to all interactions with Kodely
  • Mode-specific instructions: Only apply when using a particular mode

Custom Modes

Custom modes allow you to create specialized AI personas for specific tasks or domains.

Creating Custom Modes

To create a custom mode:

  1. Open the Kodely panel
  2. Click on the Settings icon (gear)
  3. Select "Custom Modes"
  4. Click "Create New Mode"
  5. Enter a name for the mode
  6. Write a detailed description of the mode's purpose and behavior
  7. Configure any additional settings
  8. Click "Save"

Example Custom Modes

  • Security Auditor: Specialized in finding security vulnerabilities in your code
  • Performance Optimizer: Focused on identifying and fixing performance bottlenecks
  • Documentation Writer: Specialized in creating clear and comprehensive documentation
  • Test Engineer: Focused on writing effective tests and improving test coverage
  • UI/UX Specialist: Specialized in frontend development and user experience

Writing Effective Mode Descriptions

When creating a custom mode, consider including:

  • The mode's primary purpose and focus areas
  • Specific expertise or knowledge the mode should have
  • How the mode should approach problems
  • Any specific frameworks, libraries, or tools the mode should be familiar with
  • The tone and style of communication the mode should use

Model Context Protocol (MCP)

Model Context Protocol (MCP) is a framework that allows Kodely to use external tools and services, extending its capabilities beyond what's built in.

What is MCP?

MCP is a protocol that defines how Kodely can:

  • Discover available tools
  • Understand tool capabilities and parameters
  • Call tools to perform actions
  • Process the results from tool calls

Benefits of MCP

  • Extensibility: Add new capabilities to Kodely without waiting for official updates
  • Integration: Connect Kodely to your existing tools and services
  • Customization: Create tools specific to your workflow or domain
  • Automation: Build complex workflows that combine multiple tools

Example Use Cases

  • Database integration: Allow Kodely to query your database
  • API testing: Create tools to test API endpoints
  • Deployment: Build tools to deploy your application
  • Code analysis: Integrate with static analysis tools
  • Documentation generation: Create tools to generate documentation from code

For more information on MCP, including how to create your own tools, see theMCP documentation.

Next Steps

Now that you understand Kodely's features, you can explore more advanced usage: