Home›Blog›Agentic AI System Designs
Agentic AI System Designs · The map

Four Ways to Build an Agent

Agent frameworks, harnesses, runtimes and builders solve different problems, and choosing the wrong kind is the costliest mistake in an agent project. This map explains the four, helps you pick, and opens a series that takes the main systems apart one by one.

FL
FrontierAGI Team
⚠️ How to read this (as of 7 Nov 2026)The four-way framing is adapted from a third-party infographic (Denis Panjuta); product placements are our reading of each project's own description and may change. Disclosure: Claude, made by Anthropic, helps write this site, and this series covers Anthropic, Google and OpenAI products.

1. Why this series

2. The four kinds

🗺️ The four kinds of agent system (tap one)

Framing adapted from an infographic by Denis Panjuta (third party); descriptions and examples checked against each project where we could.

3. Which kind do you need?

🧭 Which kind do you need?

A rule of thumb, not a law. Real systems often combine kinds: a framework agent deployed on a runtime, or a harness built with a framework.

4. How they fit together

🥞 How the kinds stack

Buildera visual layer for people who do not code; often runs its own runtime
Harnessa working agent; can be built with a framework (Deep Agents is built on LangGraph)
Frameworkthe parts: models, tools, state, flow
Runtimeruns any of the above in production: scaling, sessions, identity, tracing

Our simplification. The Agent Layer series covers the infrastructure underneath all four.

5. The eight layers inside every harness

🏛️ The reference stack used in every teardown

L1 🖥️ InterfacesWhere people (or other agents) talk to it
L2 🧭 Sessions & orchestrationWho owns a task: sessions, background tasks, subagents
L3 🔁 Agent loopHow it decides the next step
L4 🧠 Context engineWhat the model sees on each turn
L5 🛠️ ToolsWhat it can do
L6 🛡️ Policy & safetyWhat it may do
L7 📦 ExecutionWhere actions actually run
L8 🧬 Model & telemetryWhich model, and what gets logged

Each harness article draws its real components into these eight bands and traces one task through them, so you can compare systems layer by layer.

6. What is published, and what is next

7. The series

Agentic AI System Designs: four kinds of agent system, one article each

Start here🧭 Four ways to build an agent (this article)Harness🧰 Claude CodeHarness🧰 CodexHarness🧰 Gemini CLI, Antigravity & JulesHarness🧰 Deep AgentsHarness🧰 Hermes AgentFramework🧱 OpenAI Agents SDK · comingFramework🧱 Google ADK · comingFramework🧱 LangChain / LangGraph · comingFramework🧱 CrewAI · comingFramework🧱 Claude Agent SDK & Microsoft Agent Framework · comingRuntime⚙️ Amazon Bedrock AgentCore · comingRuntime⚙️ LangSmith Deployment · comingRuntime⚙️ Temporal · comingRuntime⚙️ Vertex AI Agent Engine & Azure AI Foundry · comingBuilder🧩 n8n · comingBuilder🧩 Zapier Agents · comingBuilder🧩 Microsoft Copilot Studio · comingBuilder🧩 Make · coming

See them all on one page: Agentic Systems at a Glance.

8. Sources