Agent Architecture
Four components that make an AI agent work
LLM Core
The reasoning engine that plans, decides, and generates responses
Brain: processes language and makes decisions
Tools
APIs, databases, browsers: the hands the agent uses to act
Hands: executes actions in the real world
Memory
Short-term (conversation) and long-term (vector DB) context
Memory: maintains continuity across sessions
Orchestration
Planning loops, task decomposition, and self-reflection
Manager: coordinates multi-step workflows
To understand the Agentic Web, you need to know its building blocks. This module explores the key components that enable agents to function, from their internal "minds" to the external tools they use to interact with the world.
Agent Tool Ecosystem
How agents connect to real-world services across domains
Search
Perplexity
Genspark
You.com
Code
Cursor
Claude Code
Antigravity
Data
MCP Sheets
Postgres MCP
Browser MCP
Commerce
Shopify Agents
Stripe MCP
Payment APIs