PROVIDERS

Configure AI model providers for use with Turbine.

§ CLAUDE (ANTHROPIC)

Claude is the recommended primary provider for code generation tasks.

Setup

turbine config set claude.apiKey YOUR_API_KEY

Configuration

"claude": {
  "priority": 1,
  "rateLimit": 1000,
  "model": "claude-3-opus",
  "maxTokens": 4096
}

Rate limits: 1000 RPM (requests per minute) on standard tier.

§ CODEX (OPENAI)

OpenAI Codex for code completion and generation.

Setup

turbine config set codex.apiKey YOUR_API_KEY

Configuration

"codex": {
  "priority": 2,
  "rateLimit": 500,
  "model": "gpt-4-turbo",
  "maxTokens": 4096
}

Rate limits vary by tier. Check OpenAI dashboard for your limits.

§ GEMINI (GOOGLE)

Google Gemini for cost-effective background tasks.

Setup

turbine config set gemini.apiKey YOUR_API_KEY

Configuration

"gemini": {
  "priority": 3,
  "rateLimit": 2000,
  "model": "gemini-pro",
  "maxTokens": 8192
}

Gemini offers higher rate limits and lower costs for suitable tasks.

§ CUSTOM PROVIDERS

Add custom or self-hosted model providers:

"custom": {
  "priority": 4,
  "endpoint": "https://your-api.com/v1",
  "apiKey": "YOUR_KEY",
  "model": "custom-model",
  "rateLimit": 100
}

Custom providers must implement the OpenAI-compatible API format.

§ PROVIDER PRIORITY

Provider priority determines routing order. Lower numbers = higher priority. When the primary provider hits rate limits, requests automatically route to the next available provider.

Priority 1: claude → Primary choice
Priority 2: codex → First fallback
Priority 3: gemini → Second fallback