Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/nullclaw/nullclaw/llms.txt

Use this file to discover all available pages before exploring further.

The nullclaw onboard command guides you through initial configuration, creating ~/.nullclaw/config.json with your provider, channels, and security settings.

​
Usage

# Quick setup with API key
nullclaw onboard --api-key sk-or-... --provider openrouter

# Interactive wizard
nullclaw onboard --interactive

# Reconfigure channels only
nullclaw onboard --channels-only

​
Options

​
--api-key
string
API key for the provider (quick mode)
​
--provider
string
Provider name: openrouter, openai, anthropic, groq, ollama, etc.
​
--interactive
boolean
Run full interactive wizard (recommended for first-time setup)
​
--channels-only
boolean
Reconfigure channels and allowlists only (preserve existing provider config)

​
Quick Mode

Fastest way to get started: 
nullclaw onboard --api-key sk-or-v1-abc123 --provider openrouter
This creates a minimal config: 
{
  "models": {
    "providers": {
      "openrouter": {
        "api_key": "sk-or-v1-abc123"
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "openrouter/anthropic/claude-sonnet-4"
      }
    }
  }
}

​
Interactive Wizard

The full wizard walks through:
1

Provider selection

Choose from 22+ AI providers:
  • OpenRouter (recommended - aggregates 100+ models)
  • Anthropic (Claude)
  • OpenAI (GPT-4, o1)
  • Groq (fast inference)
  • Ollama (local models)
  • And more…
2

API key setup

Enter your API key. The wizard will:
  • Validate the key format
  • Test connectivity
  • Encrypt the key with ChaCha20-Poly1305
3

Default model selection

Pick your primary model:
  • OpenRouter: openrouter/anthropic/claude-sonnet-4
  • Anthropic: claude-4
  • OpenAI: gpt-4o, o1
  • Ollama: llama3.2, qwen2.5
4

Channel configuration

Enable and configure messaging channels:
  • Telegram bot (token + allowlist)
  • Discord bot (token + guild)
  • Signal (phone number)
  • Nostr (keypair generation)
  • IRC, Slack, Matrix, etc.
5

Memory backend

Choose memory storage:
  • SQLite (recommended - hybrid vector + FTS5)
  • Markdown (simple file-based)
  • PostgreSQL (for production)
  • Redis (caching layer)
6

Security settings

Configure autonomy level:
  • Supervised (requires approval for risky commands)
  • Autonomous (full automation)
  • Restricted (read-only)
Set workspace restrictions:
  • workspace_only: true (recommended)
  • Command allowlist
  • Resource limits
7

Nostr setup (optional)

If enabling Nostr:
  • Generate new keypair or import existing
  • Configure relays (wss://relay.damus.io, etc.)
  • Set owner pubkey for DM allowlist
  • Install nak CLI tool

​
Channels-Only Mode

Reconfigure channels without touching provider settings: 
nullclaw onboard --channels-only
Use this to:
  • Add new channels
  • Update allowlists
  • Change channel credentials
  • Enable/disable channels

​
Examples

​
OpenRouter + Telegram

nullclaw onboard --interactive

# Wizard prompts:
# 1. Provider: openrouter
# 2. API key: sk-or-v1-...
# 3. Model: openrouter/anthropic/claude-sonnet-4
# 4. Enable Telegram? yes
# 5. Bot token: 123456:ABC-DEF...
# 6. Allowed users: your_telegram_username
Result: 
{
  "models": {
    "providers": {
      "openrouter": {"api_key": "enc2:..."}
    }
  },
  "agents": {
    "defaults": {
      "model": {"primary": "openrouter/anthropic/claude-sonnet-4"}
    }
  },
  "channels": {
    "telegram": {
      "accounts": {
        "main": {
          "bot_token": "enc2:...",
          "allow_from": ["your_telegram_username"]
        }
      }
    }
  }
}

​
Local Ollama

nullclaw onboard --api-key none --provider ollama

# Model: llama3.2
# Base URL: http://localhost:11434 (default)

​
Configuration Output

Onboarding creates ~/.nullclaw/config.json with:
  • Encrypted API keys (ChaCha20-Poly1305)
  • Provider and model selection
  • Channel credentials and allowlists
  • Memory backend configuration
  • Security policies
  • Gateway settings
You can manually edit ~/.nullclaw/config.json after onboarding. See Configuration for full reference.

​
Validation

After onboarding, verify your setup: 
# Check configuration
nullclaw status

# Run diagnostics
nullclaw doctor

# Test agent
nullclaw agent -m "Hello!"

​
See Also