Architect Documentation

Documentation, integration guides, and examples for Architect — a powerful CLI for AI agents and software engineers.

Contents

Core

How Architect works — configuration, tools, agents, skills, and the API server.

• Interactive Mode — TUI: keyboard shortcuts, commands, modes, checkpoints
• Configuration — Settings reference (settings.json)
• Providers — LLM provider setup (OpenAI, Anthropic, Ollama, ZAI)
• Tools — Tool catalog, parameters, and permission system
• Agents — Create and configure AI agents
• Skills — Inject domain knowledge into context
• Memory — Persistent knowledge across sessions
• Mentions — Inline references (@file, @url, @git)
• MCP Servers — Connect external tool servers via Model Context Protocol
• Security — Authentication, CORS, jail, Landlock, SSRF protection
• Headless Mode — Non-interactive CLI for automation
• API Server — HTTP/WebSocket server for external clients

Integrations

Complete guides for connecting Architect to external services.

• Telegram Bot — Chat with Architect through Telegram

Examples

Ready-to-use agent and skill definitions.

• Agents — Community agent definitions
• Skills — Community skill definitions

Spec

Open standards.

• AGENTS.md specification (coming soon)