How to Customize the AI Model Used in the AI Website Cloner

The AI Website Cloner supports multiple LLMs through skill configuration files and environment variables, allowing you to swap between Claude, OpenAI, Gemini, or other providers without modifying core cloning logic.

The JCodesMore/ai-website-cloner-template repository does not hard-code a single LLM for its website cloning agents. Instead, it exposes model selection through external configuration files, making it straightforward to customize the AI model used in the website cloner based on your specific needs or API access.

Understanding Model Configuration

The repository uses a skill-based architecture where each supported agent reads its model configuration from platform-specific skill files when the /clone-website command is invoked.

Skill Configuration Files

The model identifier is defined in platform-specific skill files rather than in the core codebase:

• AGENTS.md – Central description of supported agents and default settings
• .claude/skills/clone-website/SKILL.md – Claude Code skill definition containing the model field (e.g., claude-3-opus-20240229)
• .github/skills/clone-website/SKILL.md – GitHub Copilot/Continue skill configuration
• .gemini/commands/clone-website.toml – Gemini CLI skill configuration

When you customize the AI model used in the website cloner, you modify these files to point to your preferred LLM identifier.

Environment Variable Overrides

For runtime flexibility, the repository supports environment variable overrides that take precedence over static configuration files:

• OPENAI_MODEL – Overrides the model for OpenAI-based agents
• GEMINI_MODEL – Overrides the model for Gemini-based agents

The repository ships with a .env.example file that you can copy to .env and populate with your preferred model settings.

Step-by-Step Customization Guide

Method 1: Edit Skill Files Directly

To permanently change the model for a specific agent, edit its skill configuration:

# For Claude Code users

nano .claude/skills/clone-website/SKILL.md

# Replace the model line

# Old: model: claude-3-opus-20240229

# New: model: claude-3-sonnet-20240229

Repeat this process for .github/skills/clone-website/SKILL.md or .gemini/commands/clone-website.toml depending on your agent.

Method 2: Use Environment Variables

For temporary or deployment-specific overrides without editing files:

# Copy the environment template

cp .env.example .env

# Edit .env and set your preferred model

echo "OPENAI_MODEL=gpt-4o" >> .env

This method lets you customize the AI model used in the website cloner without touching version-controlled configuration files.

Method 3: Regenerate Configuration Files

After editing AGENTS.md or individual skill files, propagate changes across all platform-specific copies:

# Update AGENTS.md copies across platforms

bash scripts/sync-agent-rules.sh

# Regenerate all skill files for every supported agent

node scripts/sync-skills.mjs

Running these scripts ensures consistency across .claude/skills/, .github/skills/, and other platform directories.

Supported Model Providers

You can customize the AI model used in the website cloner to use any of the following providers:

• Anthropic Claude – Configure via .claude/skills/clone-website/SKILL.md using model IDs like claude-3-opus-20240229 or claude-3-sonnet-20240229
• OpenAI – Set via environment variable OPENAI_MODEL or in the appropriate skill file using IDs like gpt-4o
• Google Gemini – Configure in .gemini/commands/clone-website.toml or via GEMINI_MODEL environment variable

Because the model is defined outside the core cloning logic in AGENTS.md and platform-specific skill files, you can freely migrate between providers or upgrade to newer models as they become available.

Summary

• The AI Website Cloner uses skill configuration files rather than hard-coded model selections
• Edit .claude/skills/clone-website/SKILL.md, .github/skills/clone-website/SKILL.md, or .gemini/commands/clone-website.toml to change models permanently
• Use OPENAI_MODEL or GEMINI_MODEL environment variables for runtime overrides via .env
• Run scripts/sync-agent-rules.sh and scripts/sync-skills.mjs after editing configuration files to regenerate platform-specific copies
• The modular architecture allows swapping between Claude, OpenAI, Gemini, or other supported providers without modifying the core cloning codebase

Frequently Asked Questions

What is the default model used by the AI Website Cloner?

The default model varies by agent and platform. According to the AGENTS.md file and platform-specific skill files, Claude Code typically defaults to claude-3-opus-20240229, while OpenAI and Gemini configurations reference their respective latest stable models. Check the specific skill file for your agent to confirm the current default.

Can I use a local or custom fine-tuned model?

Yes, provided your agent supports it. Since you customize the AI model used in the website cloner by editing the model identifier string in skill files or environment variables, you can specify any model identifier your API key has access to, including fine-tuned models or local endpoints that are compatible with the OpenAI, Claude, or Gemini APIs.

Do I need to restart the agent after changing the model?

No restart is required for environment variable changes if you export them before running the agent. For skill file edits, the changes are picked up automatically the next time you invoke the /clone-website command. However, if you edit AGENTS.md or want to ensure consistency across platforms, run bash scripts/sync-agent-rules.sh and node scripts/sync-skills.mjs before executing the command.

Where should I store my API keys and model preferences?

Store sensitive configuration like API keys and model overrides in a .env file in the repository root. The repository provides a .env.example template that you can copy and customize. Never commit actual API keys to version control; the .env file should be listed in .gitignore by default.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}