Models & Providers

Configure AI models in your preferred tool to use Aexol.ai's built-in Spectral model or other supported providers.

Aexol.ai Built-in Model

The built-in/spectral model is Aexol's native AI, optimized for understanding and generating Aexol specifications.

Endpoint: https://api.aexol.ai/v1
Model name: built-in/spectral
Auth: Team API key (sk-aexol-team-...)

Get your team API key from the Studio → Team settings.

OpenCode

Add a provider entry in your opencode.json or opencode.jsonc:

{
  "provider": {
    "aexol.ai": {
      "name": "Aexol AI",
      "options": {
        "baseURL": "https://api.aexol.ai/v1",
        "apiKey": "sk-aexol-team-..."
      },
      "models": {
        "built-in/spectral": {
          "name": "built-in/spectral"
        }
      }
    }
  }
}

Config locations:

• Global: ~/.config/opencode/opencode.json
• Project-level: opencode.json in your project root

OpenCode resolves the model as {baseURL}/models/{modelName}/v1/chat/completions, so built-in/spectral maps to https://api.aexol.ai/v1/models/built-in/spectral/v1/chat/completions.

Also connect MCP (optional)

To use Aexol tools (validate, generate, etc.) alongside the model, also add the MCP config:

{
  "mcp": {
    "aexol": {
      "type": "remote",
      "url": "https://api.aexol.ai/mcp",
      "headers": {
        "Authorization": "Bearer sk-aexol-team-..."
      },
      "oauth": false
    }
  }
}

Claude Code

# Add Aexol as an OpenAI-compatible provider
claude config set --global model aexol/built-in/spectral

Or configure directly in Claude's config with the Aexol.ai base URL and your team API key.

See Working with AI & IDEs for Claude Code MCP setup.

Cursor IDE

In Cursor settings, add a custom model provider:

• Provider type: OpenAI-compatible
• Base URL: https://api.aexol.ai/v1
• API Key: sk-aexol-team-...
• Model: built-in/spectral

Windsurf / Codeium

In Windsurf settings, configure a custom AI provider:

• Base URL: https://api.aexol.ai/v1
• API Key: sk-aexol-team-...
• Model: built-in/spectral

Continue (VS Code Extension)

In ~/.continue/config.json:

{
  "models": [
    {
      "title": "Aexol Spectral",
      "provider": "openai",
      "model": "built-in/spectral",
      "apiBase": "https://api.aexol.ai/v1",
      "apiKey": "sk-aexol-team-..."
    }
  ]
}

Other OpenAI-compatible clients

Any client that supports custom OpenAI-compatible endpoints works:

Getting a Team API Key

1. Open Aexol Studio
2. Go to Team settings → API Keys
3. Click Create new key
4. Copy the key — it starts with sk-aexol-team-

Next Steps

• Set up MCP tools → Working with AI & IDEs
• Configure remote MCP endpoint → Remote MCP
• Run inference → CLI Commands