Nautilus AI
Nautilus AI is the core intelligence layer of the Green Monkey Dashboard, offering two distinct modes for interacting with language models.
Dual-Mode Architecture
Nautilus provides a tabbed interface with two modes:
Direct LLM Mode
Send prompts directly to your configured language model with no middleware. This gives you raw, unfiltered access to the AI.
- Choose from multiple providers (Anthropic Claude, OpenAI, OpenRouter, Ollama)
- Select specific models per provider
- Adjust temperature and token limits
- View raw responses in real time
Gateway Mode
Route prompts through the Green Monkey agent gateway for orchestrated, multi-step interactions. The gateway adds:
- Persistent memory across sessions
- Tool access (web search, file operations, etc.)
- Behavioral guidelines from your
SOUL.md - Context from your
USER.mdandIDENTITY.md
Supported Providers
| Provider | Models | Notes |
|---|---|---|
| Anthropic | Claude 4, Claude 3.5 Sonnet | Best for reasoning and long context |
| OpenAI | GPT-4o, GPT-4, GPT-3.5 Turbo | Wide ecosystem support |
| OpenRouter | 100+ models | Single API key for multiple providers |
| Ollama | Llama, Mistral, etc. | Run locally, no API key needed |
| Custom | Any OpenAI-compatible API | For self-hosted or niche providers |
Configuration
- Navigate to the LLM Connection tab in the dashboard
- Select your provider
- Enter your API key
- Choose a default model
- Click Test Connection to verify
The configuration is saved to LLM_CONFIG.md in your workspace.
Usage Tips
- Use Direct Mode when you want fast, simple completions
- Use Gateway Mode when you need the agentās full capabilities (memory, tools, personality)
- Switch between modes using the tab interface at the top of the Nautilus panel
Last updated on