# hubagenticai — the Agentic AI Hub > The definitive Agentic AI Hub and Compare GenAI resource. Step-by-step Model Context Protocol (MCP) and AI agent tutorials, honest model comparisons (Claude vs GPT vs Gemini), governance depth, and battle-tested production templates. Evergreen, multi-vendor (Claude, GPT, Gemini, open-weights) guides to building AI agents: MCP servers, agent loops, multi-agent orchestration, memory, evals, security, observability, cost control and governance. Organized as three ordered learning tracks. ## Beginner track New to agents, or used ChatGPT but never built one. Understand what an agent really is, then build one that works — a tool server, an agent loop, and an agentic coding workflow. - [How an AI agent actually works: build the loop behind every framework](https://hubagenticai.com/tutorials/how-an-agent-actually-works/): Tutorial. Every agent framework wraps the same ~40-line loop: model decides, tool runs, result feeds back. Build it in plain Python and the whole ecosystem stops being magic. - [The agentic AI tech stack, layer by layer: what you actually need to build agents](https://hubagenticai.com/compare/agentic-ai-tech-stack/): Decision guide. Seven layers make up every serious agentic system — models, serving, harness, protocols, knowledge, evals, and controls. What each layer does, the main options in each, and where the choices actually matter. - [MCP vs. A2A vs. AG-UI: the agent protocol stack, explained with one diagram](https://hubagenticai.com/tutorials/agent-protocol-stack-mcp-a2a-ag-ui/): Tutorial. The three agent protocols solve three different problems — agent-to-tools, agent-to-agent, and agent-to-user. Learn which layer each owns and when you need more than one. - [How to build your first MCP server, step by step (Python and TypeScript)](https://hubagenticai.com/tutorials/build-your-first-mcp-server/): Tutorial. Build a working Model Context Protocol server with tools and resources, connect it to a real client, and understand what actually happens on the wire. - [Claude Code from zero: an agentic coding workflow, step by step](https://hubagenticai.com/tutorials/claude-code-from-zero/): Tutorial. Install Claude Code, learn the core loop, and build up to a disciplined agentic workflow: CLAUDE.md, plan mode, custom skills, hooks, and MCP servers. - [How to write Agent Skills (SKILL.md) that work across coding agents](https://hubagenticai.com/tutorials/write-agent-skills/): Tutorial. A skill is a folder with a SKILL.md file that a coding agent loads only when it needs it. Build a working release-notes skill with a bundled script, put it where Claude Code, Codex, Copilot, Cursor and Gemini CLI can all find it, and fix it when it does not trigger. - [How to choose an LLM for your agent: Claude, GPT, Gemini, or open-weights](https://hubagenticai.com/compare/choose-an-llm-for-agents/): Decision guide. General benchmarks predict agent performance poorly. The seven criteria that matter for agentic workloads, a tiering strategy that beats single-model loyalty, and the eval-driven way to decide. - [How to build a personal AI agent for daily productivity (note and task triage)](https://hubagenticai.com/tutorials/personal-productivity-agent/): Tutorial. Build an agent that triages your notes into tasks, references, and archives — runnable today with zero API keys, upgradeable to any LLM, and scalable from personal to team use. ## Intermediate track You’ve built a demo agent and want it to be reliable. Structured outputs, real integrations, memory, multi-agent orchestration, and evals — the engineering that turns a demo into a system. - [Structured outputs for agents: make models emit JSON you can actually trust](https://hubagenticai.com/tutorials/structured-outputs-for-agents/): Tutorial. Agents break where prose meets code. Build the validate-and-retry loop that turns flaky model JSON into a reliable contract — in plain Python, runnable without an API key. - [How to wrap a REST API in an MCP server (without handing the agent the whole API)](https://hubagenticai.com/tutorials/wrap-a-rest-api-in-mcp/): Tutorial. Turn any REST API into agent-safe MCP tools: a narrow allowlist instead of a proxy, PII redaction before the model sees data, and a fake API so you can test the whole thing offline. - [How to migrate an MCP server to the stateless 2026 spec (mcp v2 and TypeScript SDK v2)](https://hubagenticai.com/tutorials/migrate-mcp-server-to-2026-spec/): Tutorial. `pip install mcp` now gives you mcp 2.x and your FastMCP import breaks. Move a working MCP server to the stateless 2026-07-28 spec in Python and TypeScript, without locking out the clients that still use the old protocol. - [MCP servers vs. direct CLI calls: when each wins (with measured token costs)](https://hubagenticai.com/compare/mcp-vs-direct-cli-calls/): Decision guide. Agents can reach tools through an MCP server or by running CLI commands directly. A decision framework, real token counts for four MCP servers and their CLI equivalents, and what tool search changes. - [Skills vs. tools vs. hooks vs. subagents vs. prompts: when to use what](https://hubagenticai.com/compare/skills-tools-hooks-subagents/): Decision guide. Agentic coding assistants offer half a dozen extension mechanisms, and picking wrong wastes context or bites silently. One decision table for instructions files, skills (which now include slash commands), MCP tools, hooks, and subagents. - [Agent memory patterns: vector stores, knowledge graphs, and context engineering](https://hubagenticai.com/tutorials/agent-memory-patterns/): Tutorial. Why agents forget, the four memory patterns that fix it, and a working retrieval loop you can run in plain Python before you buy a vector database. - [RAG vs. agent memory vs. fine-tuning: which knowledge problem do you actually have?](https://hubagenticai.com/compare/rag-vs-memory-vs-fine-tuning/): Decision guide. Three ways to make a model "know" things, routinely confused with each other. A diagnostic for picking the right one — and the order to try them in. - [Multi-agent orchestration explained: build an orchestrator and sub-agents from scratch](https://hubagenticai.com/tutorials/multi-agent-orchestration-explained/): Tutorial. Build the agent harness pattern in ~80 lines of plain Python — an orchestrator that plans, delegates to specialist sub-agents, and synthesizes — with a mock model so you can run it instantly. - [Single agent vs. multi-agent: when splitting actually helps](https://hubagenticai.com/compare/single-agent-vs-multi-agent/): Decision guide. Multi-agent architectures are fashionable and usually premature. The three forces that genuinely justify a split, the costs nobody mentions, and a staged migration path. - [Agent frameworks vs. rolling your own harness: build or buy the loop?](https://hubagenticai.com/compare/agent-framework-vs-own-harness/): Decision guide. LangGraph, CrewAI, Microsoft Agent Framework, Pydantic AI, the OpenAI, Google and Claude agent SDKs — or 200 lines of your own code? A decision framework based on what agent frameworks actually provide, plus a dated map of the options. - [How to evaluate an AI agent: build a golden-task eval harness from scratch](https://hubagenticai.com/tutorials/evaluate-your-agent/): Tutorial. Agents without evals break silently on every prompt tweak. Build a small golden-task harness in plain Python — tasks, checks, scoring, and a regression gate you can run in CI. - [How to secure your agent supply chain: pinning, provenance, and MCP server vetting](https://hubagenticai.com/enterprise/agent-supply-chain-security/): Playbook. A hands-on guide to locking down the packages and MCP servers your agents run on: hash-pinned lockfiles, release cooldowns, provenance checks, .pth audits, SBOMs, and tool-definition pinning against MCP rug pulls. Every command tested. ## Advanced track You run agents for a team or company — or will soon. Security, human oversight, observability, cost control, and governance for agents doing real work in production. - [Workflow vs. agent: when BPMN-style orchestration beats dynamic reasoning](https://hubagenticai.com/compare/workflow-vs-agent/): Decision guide. The most expensive architecture mistake in enterprise AI is putting an agent where a workflow belongs — or vice versa. A decision framework with a bright-line design principle. - [Prompt injection for agentic systems: a working threat model](https://hubagenticai.com/enterprise/prompt-injection-threat-model/): Playbook. When agents read email, web pages, and documents, every input is a potential instruction. The lethal trifecta, the controls that actually work, and the ones that only feel like they do. - [Human-in-the-loop patterns that scale: approval design for agent actions](https://hubagenticai.com/enterprise/human-in-the-loop-patterns/): Playbook. Naive HITL either rubber-stamps everything or drowns reviewers. Five approval patterns, a materiality matrix for choosing, and the metrics that tell you when to loosen the loop. - [Agent observability: what to log, what to alert on, and what to keep](https://hubagenticai.com/enterprise/agent-observability/): Playbook. You cannot debug, defend, or bill what you did not record. The trajectory-log specification, the four signals worth alerting on, and retention that satisfies both engineers and auditors. - [How to trace AI agents with OpenTelemetry (agent, LLM, and tool spans)](https://hubagenticai.com/tutorials/trace-agents-with-opentelemetry/): Tutorial. Instrument a small Python agent loop with the OpenTelemetry GenAI semantic conventions: one invoke_agent span per run, a chat span per model call, an execute_tool span per tool call, plus token usage and errors. Then use the trace to explain what a run cost and where its time went. - [Run agents locally: LLM serving plus MCP on your own hardware](https://hubagenticai.com/tutorials/run-agents-locally/): Tutorial. Serve an open-weights model with llama.cpp on your own machine (ARM64 included), expose an OpenAI-compatible endpoint, and wire a local agent loop to MCP tools — no cloud required. - [Long-running agent tasks: MCP Tasks vs. A2A tasks vs. durable execution](https://hubagenticai.com/compare/long-running-agent-tasks/): Decision guide. When agent work takes minutes, hours, or a human approval, three different things called "tasks" compete for the job. What each one guarantees, what it does not, and how they fit together. - [FinOps for AI agents: metering token spend before it meters you](https://hubagenticai.com/enterprise/finops-for-agents/): Playbook. An architecture for agent cost governance: OpenTelemetry GenAI conventions for instrumentation, a gateway for enforcement, and the reporting dimensions finance will actually ask for. - [The agent adoption ladder: personal → team → enterprise, without the faceplant](https://hubagenticai.com/enterprise/agent-adoption-ladder/): Playbook. What actually changes when an agent graduates from one laptop to a team to a company — identity, secrets, review, cost — and the checklist for each rung. - [AI agent governance for regulated industries: a practical framework](https://hubagenticai.com/enterprise/agent-governance-regulated-industries/): Playbook. A working control framework for agentic AI in banks, insurers and other regulated organizations: five control layers, a worked example of a refund agent, and a mapping table to SR 26-2 concepts and EU AI Act articles. - [SR 26-2 explained: governing agentic AI when model-risk guidance excludes it](https://hubagenticai.com/enterprise/sr-26-2-agentic-ai-governance/): Playbook. SR 26-2 replaced SR 11-7 in April 2026 and explicitly puts generative and agentic AI out of scope. What the guidance actually says, what that exclusion does and does not mean, and a practical control framework for the agents that fall into the gap. - [EU AI Act timeline for AI agents after the Digital Omnibus (2026–2028)](https://hubagenticai.com/enterprise/eu-ai-act-timeline-for-agents/): Playbook. Which EU AI Act obligations already apply to AI agents, which moved under the Digital Omnibus (Regulation (EU) 2026/1744), and a checklist of what an agent team should do this quarter. ## Compare GenAI — Decision Frameworks (https://hubagenticai.com/compare) > comparegenai.com redirects to https://hubagenticai.com/compare/ for unbiased, head-to-head model and architecture verdicts. - [How to choose an LLM for your agent](https://hubagenticai.com/compare/choose-an-llm-for-agents/): Claude, GPT-4o, Gemini, or open-weights evaluated on tool-calling reliability and error recovery. - [Agent framework vs own harness](https://hubagenticai.com/compare/agent-framework-vs-own-harness/): When to write a deterministic loop vs adopting LangGraph or CrewAI. - [RAG vs agent memory vs fine-tuning](https://hubagenticai.com/compare/rag-vs-memory-vs-fine-tuning/): Operational boundaries and cost profiles. - [Workflow vs agent](https://hubagenticai.com/compare/workflow-vs-agent/): When deterministic DAGs beat reasoning loops. - [Single-agent vs multi-agent](https://hubagenticai.com/compare/single-agent-vs-multi-agent/): When multi-agent decomposition actually helps. - [The agentic AI tech stack](https://hubagenticai.com/compare/agentic-ai-tech-stack/): The 7 layers from foundation model to audit storage. - [MCP vs direct CLI tool execution](https://hubagenticai.com/compare/mcp-vs-direct-cli-calls/): When protocol standardization beats raw child processes. - [Skills vs tools vs hooks vs subagents](https://hubagenticai.com/compare/skills-tools-hooks-subagents/): Clarifying extension taxonomy for AI systems. - [Long-running agent tasks](https://hubagenticai.com/compare/long-running-agent-tasks/): Checkpoints, idempotency, and human resumption. ## Reference & Tools - [Start here](https://hubagenticai.com/start/): goal picker and how to use the hub - [Latest Agentic AI Radar](https://hubagenticai.com/radar/): quarterly adopt/trial/assess/hold map of agentic AI tools, frameworks, and protocols (latestagenticai.com) - [Templates](https://hubagenticai.com/templates/): copy-paste starters (MCP servers, eval harnesses, prompts, checklists) - [Glossary](https://hubagenticai.com/glossary/): agentic AI terms in plain language - [FAQ](https://hubagenticai.com/faq/) - [Resources](https://hubagenticai.com/resources/): annotated official specs and docs ## Companion Engineering Portals - [MCP Step by Step](https://mcpstepbystep.com): The dedicated specification field manual for Model Context Protocol (2026 Streamable HTTP). - [Build REST API](https://buildrestapi.com): Comprehensive guide to REST API architecture, OpenAPI 3.1 contracts, and tool calling for AI agents.