1. Connecting to Cursor IDE

Cursor supports custom OpenAI-compatible endpoints directly through its settings menu:

1. Open Cursor Settings (Gear icon in top right, or Ctrl/Cmd + Shift + J).
2. Navigate to Models → OpenAI API Key.
3. Paste your key: xpl_YOUR_40_HEX_KEY (do NOT add "Bearer ").
4. Toggle ON Override OpenAI Base URL and enter:
   https://api.experientiallabs.ai/v1
5. Under the Model Names list, click + Add Model and add these two custom slugs:
   • gpt-6-astra
   • claude-fable-5.1
6. Open Cursor Chat (Cmd/Ctrl + L), select gpt-6-astra from the model dropdown, and enjoy coding with 1.05M tokens of free reasoning!

2. Connecting to Claude Code CLI (Native Anthropic Tool)

Because Experiential Labs features a transparent Anthropic Messages translator, you can even drive OpenAI's GPT-6 Astra using Anthropic's official claude CLI tool!

Add this block to your ~/.zshrc, ~/.bashrc, or launch shell:

# Point Claude Code to Experiential Labs gateway
export ANTHROPIC_BASE_URL="https://api.experientiallabs.ai"
export ANTHROPIC_API_KEY="xpl_YOUR_API_KEY_HERE"

# Select either model:
export ANTHROPIC_MODEL="gpt-6-astra"
# OR: export ANTHROPIC_MODEL="claude-fable-5.1"

# Prevent early compaction by declaring the full context window
export CLAUDE_CODE_MAX_CONTEXT_TOKENS=1050000

Now run claude in any project directory. Claude Code will automatically run tools, inspect project files, and refactor code using your free promotional token pool.

3. Connecting to VS Code Copilot Chat (Core 1.122+)

Modern VS Code supports custom chat language model endpoints natively without any paid Copilot subscription:

Locate your VS Code user configuration directory:

• macOS: ~/Library/Application Support/Code/User/chatLanguageModels.json
• Linux: ~/.config/Code/User/chatLanguageModels.json
• Windows: %APPDATA%\Code\User\chatLanguageModels.json

Insert the following JSON payload (notice the critical temperature: null setting):

[
  {
    "name": "Experiential Labs Free Tier",
    "vendor": "customendpoint",
    "apiType": "chat-completions",
    "models": [
      {
        "id": "gpt-6-astra",
        "name": "GPT-6 Astra (Free 1.05M)",
        "url": "https://api.experientiallabs.ai/v1/chat/completions",
        "toolCalling": true,
        "vision": true,
        "modelOptions": {
          "temperature": null,
          "top_p": null
        }
      },
      {
        "id": "claude-fable-5.1",
        "name": "Claude Fable 5.1 (Free 1M)",
        "url": "https://api.experientiallabs.ai/v1/chat/completions",
        "toolCalling": true,
        "vision": true,
        "modelOptions": {
          "temperature": 1.0
        }
      }
    ]
  }
]

In VS Code Chat, open Manage Models, right-click the Experiential Labs model, select Enter API Key, and paste your xpl_... key. VS Code stores it in your operating system's encrypted keychain.

4. Troubleshooting Common Errors