First Run

Step 1: Open The Setup Wizard

kabot config

This is the main setup surface for most users.

Step 2: Configure A Model Provider

Inside the wizard, start with Model / Auth.

Recommended beginner flow: - login to one provider first - let Kabot auto-build a primary model path - only add manual override models after the basics work

Good first providers for many users: - OpenAI - Anthropic - Gemini - Groq - OpenRouter - Ollama (Local Auto-Discovery)

Step 3: Save Configuration

Once the wizard saves: - Kabot writes config into your local runtime area - provider credentials are stored locally - later menus like Channels, Skills, and Gateway can reuse that config

Step 4: Start The Gateway

kabot gateway

Expected result: - Kabot starts listening on the configured gateway port - dashboard routes become available - channel/webhook ingress becomes available if configured

Step 5: Test CLI Chat

kabot agent -m "Hello Kabot"

This verifies: - CLI entrypoint works - your model config is usable - Kabot can build runtime context and answer

If Something Fails

Go to: - Configuration guide - Troubleshooting

And run:

kabot doctor --fix

Important Mental Model

Kabot has a few core surfaces:

• kabot config for setup and operator configuration
• kabot gateway for the web/dashboard and ingress runtime
• kabot agent for direct chat and one-shot prompts
• kabot doctor for diagnostics and smoke checks