Providers

Iron Rain supports 7 providers out of the box — both API-based and CLI-based. Use your existing subscriptions.

API vs CLI providers

API providers call model APIs directly via HTTP — you need an API key.
CLI providers spawn your existing CLI tools (Claude Code, Codex, Gemini CLI) — they use your existing subscription, no API key needed.

Multimodal support

API providers (Anthropic, OpenAI, Gemini) support image references via @./image.png or @image:path. Images are base64-encoded and sent as multimodal content. CLI and Ollama bridges receive the text-only portion.

Ollama (local, free)

Run open-source models locally. No API key, no cost, full privacy.

Setup

Install Ollama: ollama.com
Pull a model: ollama pull llama3.2
Start the server: ollama serve

Config

{
  "slots": {
    "main": { "provider": "ollama", "model": "llama3.2" }
  },
  "providers": {
    "ollama": { "apiBase": "http://localhost:11434" }
  }
}

Popular models

Model	Size	Best for
`llama3.2`	3B	Fast general tasks
`qwen2.5-coder:32b`	32B	Code generation
`deepseek-coder-v2`	16B	Code understanding
`mistral`	7B	Balanced performance

Anthropic (API)

Claude models via the Anthropic API.

Setup

Get an API key at console.anthropic.com
Set the environment variable: export ANTHROPIC_API_KEY=sk-ant-...

Config

{
  "slots": {
    "main": { "provider": "anthropic", "model": "claude-sonnet-4-20250514" }
  },
  "providers": {
    "anthropic": { "apiKey": "env:ANTHROPIC_API_KEY" }
  }
}

Available models

Model	Best for
`claude-opus-4-6`	Complex reasoning, architecture
`claude-sonnet-4-20250514`	Balanced speed/quality
`claude-haiku-4-5-20251001`	Fast, cheap tasks

Claude Code (CLI)

Use your existing Claude Code subscription — no API key needed. Iron Rain spawns claude as a subprocess.

Setup

Install Claude Code: npm install -g @anthropic-ai/claude-code
Log in: claude login

Config

{
  "slots": {
    "main": { "provider": "claude-code", "model": "claude-sonnet-4-20250514" }
  }
}

Custom binary path

If claude is not in your PATH, set apiBase to the full path of the binary.

OpenAI (API)

GPT models via the OpenAI API.

Setup

Get an API key at platform.openai.com
Set the environment variable: export OPENAI_API_KEY=sk-...

Config

{
  "slots": {
    "execute": { "provider": "openai", "model": "gpt-4o" }
  },
  "providers": {
    "openai": { "apiKey": "env:OPENAI_API_KEY" }
  }
}

Available models

Model	Best for
`gpt-4o`	Balanced, multimodal
`o3`	Advanced reasoning
`o4-mini`	Fast reasoning

Codex (CLI)

Use your OpenAI Codex subscription via the CLI.

Setup

Install Codex: npm install -g @openai/codex
Log in via the Codex CLI

Config

{
  "slots": {
    "execute": { "provider": "codex", "model": "o4-mini" }
  }
}

Gemini (API)

Google Gemini models via the Generative Language API.

Setup

Get an API key at AI Studio
Set the environment variable: export GEMINI_API_KEY=...

Config

{
  "slots": {
    "explore": { "provider": "gemini", "model": "gemini-2.5-flash" }
  },
  "providers": {
    "gemini": { "apiKey": "env:GEMINI_API_KEY" }
  }
}

Available models

Model	Best for
`gemini-2.5-pro`	Complex tasks
`gemini-2.5-flash`	Fast, balanced
`gemini-2.0-flash`	Ultra-fast

Gemini CLI

Use the Google Gemini CLI with your existing subscription.

Setup

Install the Gemini CLI
Log in via the CLI

Config

{
  "slots": {
    "explore": { "provider": "gemini-cli", "model": "gemini-2.5-flash" }
  }
}

OpenAI-compatible (any provider)

Any provider with an OpenAI-compatible API works out of the box. This includes Groq, Together AI, Fireworks, Azure OpenAI, LM Studio, and many others.

Config

{
  "slots": {
    "explore": { "provider": "groq", "model": "llama-3.3-70b-versatile" }
  },
  "providers": {
    "groq": {
      "apiKey": "env:GROQ_API_KEY",
      "apiBase": "https://api.groq.com/openai/v1"
    }
  }
}

How it works

Any provider name that isn't one of the built-in names (ollama, anthropic, claude-code, codex, gemini, gemini-cli) automatically uses the OpenAI-compatible bridge. Just provide apiKey and apiBase.

Recommended slot assignments

Some suggested configurations depending on your setup:

All-local (free, private)

{
  "slots": {
    "main":    { "provider": "ollama", "model": "llama3.2" },
    "explore": { "provider": "ollama", "model": "llama3.2" },
    "execute": { "provider": "ollama", "model": "qwen2.5-coder:32b" }
  }
}

Subscription-based (no API keys)

{
  "slots": {
    "main":    { "provider": "claude-code", "model": "claude-sonnet-4-20250514" },
    "explore": { "provider": "gemini-cli", "model": "gemini-2.5-flash" },
    "execute": { "provider": "codex",      "model": "o4-mini" }
  }
}

Power user (mixed API + local)

{
  "slots": {
    "main":    { "provider": "anthropic", "model": "claude-opus-4-6" },
    "explore": { "provider": "ollama",    "model": "qwen2.5-coder:32b" },
    "execute": { "provider": "openai",    "model": "gpt-4o" }
  }
}

Configuration

Architecture