{"slug": "level-100-claude-foundations", "title": "Level 100: Claude foundations", "summary": "Anthropic's Claude platform comprises four core components: Claude Code, an agentic command-line tool for code generation and refactoring; the Agent SDK, a framework for building multi-agent applications; the Claude API, a low-level HTTP interface for programmatic model access; and MCP, a standard for connecting Claude to external services. Each surface is designed for different use cases, from repository-level operations to scalable data extraction.", "body_md": "**There are four basic components:**\n\n**Claude Code**\n\nClaude Code is Anthropic's agentic command-line tool for code generation, refactoring, and debugging. It runs where your code lives, and you configure its behavior through a CLAUDE.md file, custom skills, and MCP integrations.\n\nThe distinction that matters: this isn't a chat window you paste code into. It's an agent that reads your files, makes edits, runs commands, and checks its own work inside your project.\n\nWhere it shines:\n\nPair programming on a real codebase\n\nCI/CD automation (think fixups, migrations, dependency bumps)\n\nCode review bots that comment on diffs\n\nMonorepo tooling where context spans many packages\n\nIf your task is \"operate on this repository,\" Claude Code is the surface built for it.\n\n**Agent SDK**\n\nThe Agent SDK is a framework for building multi-agent applications. It hands you the primitives you'd otherwise write yourself: agentic loops, subagent orchestration, tool calling, and full lifecycle hooks.\n\nThe value here is that the hard parts of agent plumbing are already solved. You describe the agents and tools; the SDK runs the loop, routes between subagents, and gives you hooks to observe or intercept each step.\n\nWhere it shines:\n\nCustomer support agents that route, look up, and resolve\n\nResearch pipelines that fan out across sources and synthesize\n\nAutonomous workflows that chain multiple decisions\n\nIf you're writing your own while loop around a model call and manually dispatching tools, the SDK is probably what you actually wanted.\n\n**Claude API**\n\nThe Claude API is the HTTP layer for programmatic access to the model. This is messages.create(), tool_use, structured JSON output, batch processing, and streaming. No agent framework, no CLI, just the model and your code.\n\nThis is the lowest-level, highest-control surface. You own the loop, the state, and the orchestration. In exchange, you get maximum flexibility and the ability to run at scale.\n\nWhere it shines:\n\nData extraction from unstructured text\n\nContent generation pipelines\n\nClassification at scale (batch processing is your friend here)\n\nIf you want a single well-defined transformation, run many times, with predictable structured output, the API is the direct path.\n\n**MCP: Model Context Protocol**\n\nMCP is a standard, not a client. It defines how Claude connects to your databases, APIs, and services through tool and resource interfaces. You write (or install) an MCP server that exposes capabilities, and any MCP-aware client, including Claude Code and the Agent SDK, can use them.\n\nThe point of MCP is that you build the connector once and reuse it everywhere. Instead of wiring a database into three different apps, you write one MCP server and every surface can talk to it.", "url": "https://wpnews.pro/news/level-100-claude-foundations", "canonical_source": "https://dev.to/kobester_nz/level-100-claude-foundations-16ln", "published_at": "2026-08-23 02:37:52+00:00", "updated_at": "2026-08-23 03:13:17.541903+00:00", "lang": "en", "topics": ["ai-agents", "developer-tools", "large-language-models", "ai-infrastructure"], "entities": ["Anthropic", "Claude Code", "Agent SDK", "Claude API", "MCP"], "alternates": {"html": "https://wpnews.pro/news/level-100-claude-foundations", "markdown": "https://wpnews.pro/news/level-100-claude-foundations.md", "text": "https://wpnews.pro/news/level-100-claude-foundations.txt", "jsonld": "https://wpnews.pro/news/level-100-claude-foundations.jsonld"}}