AI agent guides: from first try to production

Where to start?

The guides are reference pages, updated when the tools change. Each shows its level: 7 for beginners, 20 intermediate, 7 advanced. If you are new to the topic, read the getting-started guide first, then the one on agent instructions, then the security guide.

For a quick choice, the “Which AI agent?” quiz points you in the right direction in a few questions.

Getting started with AI agents

• Getting started Beginner
  Start with an AI agent by choosing a focused task, limiting its access, and checking its actions. Follow a practical setup process that keeps you in control.
• Your first LLM API Beginner
  Learn how to call an LLM API, protect your key, send your first request, adjust settings, estimate costs, and handle common errors before using it regularly.
• Agent instructions Intermediate
  Write testable AI agent instructions with system prompts, AGENTS.md and CLAUDE.md files, adaptable examples, permission rules, and a practical testing method.
• Deep research Beginner
  Deep research with AI brings sources together into a report. Choose suitable tools, check the evidence, and assess the findings before you decide or publish.

The leading agents, one by one

• OpenClaw Intermediate
  Install and configure OpenClaw with practical guidance on models, skills, messaging, permissions, and costs to keep your self-hosted AI agent under control.
• Hermes Agent Intermediate
  Hermes Agent uses tools, memory, and reusable skills to complete tasks. Learn how to install it, choose a model, manage costs, and control access and risks.
• Claude Code Intermediate
  Install Claude Code, choose subscription or API billing, configure CLAUDE.md, MCP, and subagents, and manage permissions, testing, and costs with clear checks.
• Codex CLI Intermediate
  Codex CLI reads, edits, and tests code from your terminal. Learn how to install it, configure permissions, write AGENTS.md, and keep project costs under control.
• Cursor Beginner
  Cursor combines a code editor with AI agents. Learn how its modes, rules, pricing, and permissions work, and how to review changes while staying in control.
• Using a coding agent Intermediate
  Work effectively with a coding agent: define the task, keep changes focused, review diffs, and test behavior to maintain code quality and control your budget.
• LangGraph Advanced
  LangGraph orchestrates stateful agents, tools, and human approvals. Learn how to design a graph, manage its memory, and prepare an application for deployment.
• CrewAI Intermediate
  CrewAI organizes AI agents into teams and workflows into flows. Learn how to install it, define tasks, connect tools, and deploy with practical checks in place.

Automate and connect

• n8n and AI Intermediate
  n8n connects AI models to business software. Learn to configure agents, tools, and memory, choose a hosting setup, and understand the full cost of running it.
• AI in Make and Zapier Beginner
  Make and Zapier bring AI into automation. Compare features, agents, billing, and limits to build useful workflows with clear controls and predictable costs.
• Choosing an MCP server Intermediate
  Choose an MCP server based on its source, permissions, and maintenance. Compare local and remote hosting, then check access controls before connecting it.
• Agent skills Intermediate
  AI agent skills provide reusable procedures. Learn how to install and create them, test their behavior, and check their permissions before using them regularly.
• Agents and messaging Intermediate
  Connect an AI agent to Telegram, WhatsApp, Slack, or Discord: compare methods, permissions, costs, and safeguards to control its access and actions safely.
• Voice agents Intermediate
  A voice agent listens, responds, and takes action over the phone or in an app. Compare architectures, latency, full operating costs, and practical safeguards.
• Computer use Intermediate
  Computer use agents operate browsers and desktops. Learn how they work, compare available tools, and manage their errors, permissions, and security risks.
• Multi-agent systems Advanced
  Multi-agent systems divide work among specialized agents. Learn when to use them, how to organize their work, and how to keep costs and failures under control.
• RAG Intermediate
  RAG connects a language model to your documents. Learn to prepare sources, split and index content, and evaluate answers to build a system you can verify.

Choosing and running a model

• Choose a model Intermediate
  Choose a language model for your tasks by comparing measured quality, actual cost, context, speed, and licensing with a practical, repeatable testing method.
• LLM API pricing Intermediate
  Understand LLM API pricing, from input and output tokens to caching and batch discounts. Estimate a realistic application budget and track actual spending.
• Local LLMs Intermediate
  Run an LLM locally: compare software, choose a model that fits your hardware, and evaluate performance, operating costs, practical limits, and everyday use.
• Quantization and GGUF Advanced
  Quantization reduces memory use for local models. Learn how GGUF, Q4 to Q8 levels, and quality trade-offs help you choose a file that fits your hardware and tasks.
• Fine-tuning vs. RAG Advanced
  Choose between RAG and fine-tuning based on your data, errors, and budget. Compare requirements, total costs, evaluation methods, and practical design examples.
• Reduce costs Intermediate
  Reduce AI agent costs by measuring successful tasks, then adjusting models, caching, context, routing, and spending limits while maintaining verified quality.

Securing and going to production

• Secure an AI agent Intermediate
  Secure an AI agent with limited permissions, protected secrets, and controlled actions. Learn how to isolate tasks, require approval, and keep useful logs.
• Agent on a VPS Advanced
  Host an AI agent on a VPS: choose a server, set up private access, protect secrets, and plan backups, recovery, and a budget that covers ongoing maintenance.
• Monitor an agent Advanced
  Monitor AI agents with useful logs, traces, and alerts. Track performance, costs, output quality, and privacy while diagnosing failures in production systems.
• Evaluate an agent Advanced
  Evaluate an AI agent's results, actions, costs, and reliability with representative tests, public benchmarks, and production monitoring to guide deployment.

In the workplace

• AI agents in business Beginner
  Deploy AI agents in your business with a defined pilot: choose use cases, manage security, governance, and costs, and measure results before scaling up safely.
• AI use policy Beginner
  An AI use policy sets approved tools, data safeguards, and review requirements. Learn what to include, with sample rules and practical steps for your rollout.
• AI Act checklist Intermediate
  Identify your AI Act obligations, classify your use cases, and gather evidence with a practical checklist for managing AI agents and tools in your business.