Live
Now Buildingericbefore.com — The Workshop·New VideoBuilding a Full-Stack App with Claude Code in 45 Min·New ResourceMCP Server Directory added to AI Tools·New ResourceCLAUDE.md Starter Template·Latest PostWhat Poker Taught Me About Product Decisions·UpdatedClaude Code resources — now 24 items·Show NotesEp. 12 — Timestamps + source code linked·Now Buildingericbefore.com — The Workshop·New VideoBuilding a Full-Stack App with Claude Code in 45 Min·New ResourceMCP Server Directory added to AI Tools·New ResourceCLAUDE.md Starter Template·Latest PostWhat Poker Taught Me About Product Decisions·UpdatedClaude Code resources — now 24 items·Show NotesEp. 12 — Timestamps + source code linked·
TALK WITH ERIC
Resources→AI Ecosystems→Anthropic / Claude

Anthropic / Claude

I build with Claude every day. The ecosystem has grown way beyond the chatbot most people know. There are coding agents, browser automation tools, a desktop app, a plugin marketplace with over 9,000 plugins, and a huge open source community building on top of all of it.

This page covers everything I've found worth knowing about. I update it regularly as things ship.

On This Page
The ModelsOfficial ProductsDocumentationFree CoursesOfficial GitHub ReposThe Plugin & Skills EcosystemCurated Community ListsCommunity Guides & Best PracticesCommunityPricing

The Models

Claude has three model tiers. Picking the right one matters.

Opus 4.6 is the most powerful. It launched February 5, 2026 with a 1 million token context window and support for agent teams that process tasks in parallel. It scores 80.9% on SWE-bench Verified. Use it for complex coding, multi-step agentic work, and anything that needs deep reasoning. API pricing is $5/$25 per million input/output tokens.

Sonnet 4.5 handles most of what Opus can do, faster and cheaper. If you don't know which model to pick, use this one.

Haiku 4.5 is the fastest. It delivers 90% of Sonnet's coding performance at 4-5x the speed. Good for high-volume tasks like customer support bots, data processing, and quick iterations.

All three are available in Claude Chat on the free tier with limits. Full model specs and API pricing live in the model documentation. Anthropic has been cutting prices hard (67% cut on Opus 4.5), so check the docs for current numbers.

Official Products

Claude Chat is where most people start. Brainstorming, analysis, writing, research. No ads. The free tier gives you access to the models with usage limits. Pro is $20/month, Max is $100/month.

Claude Code is an autonomous coding agent that runs in your terminal. It reads your entire codebase, writes and edits files, runs tests, and handles git. NASA used it for Perseverance rover path planning. It has 55k stars on GitHub and the ecosystem around it is growing fast.

Claude in Chrome is a browser extension. It lets Claude see your screen, control mouse and keyboard, and interact with web apps. Built on Anthropic's Computer Use capabilities.

Cowork is a desktop app for non-technical users. File manipulation, workflow automation, and collaboration without code. Launched January 2026.

Claude in PowerPoint lets Max, Team, and Enterprise users create presentations with AI inside Microsoft's ecosystem. Launched February 2026.

Documentation

Anthropic's docs are well written. These are the ones you'll reference most:

  • Claude Documentation - The main hub. Getting started through advanced API usage, prompt engineering, and model capabilities.
  • Claude Code Documentation - Everything specific to Claude Code. Setup, best practices, commands, hooks, skills, agents, plugins.
  • Claude Code Best Practices - How to get the most out of Claude Code. CLAUDE.md files, context management, Plan Mode, sub-agents. Written from how Anthropic's internal teams work.
  • Anthropic Learning Hub - Links to all courses, guides, and getting started materials.
  • Claude API Development Guide - Building with the API. Models, authentication, streaming, tool use, SDK setup.

Free Courses

On Anthropic Academy (anthropic.skilljar.com, free account required):

On GitHub (open source, run them yourself):

On Coursera:

Official GitHub Repos

The Plugin & Skills Ecosystem

The plugin and skills ecosystem is where Claude Code turns into something much bigger. Over 9,000 plugins, tens of thousands of community-built skills, and multiple directories to search through.

How plugins work: They bundle skills, hooks, subagents, and MCP servers into installable packages. You browse and install them from the CLI with /plugin. Skills are reusable knowledge and workflows that Claude loads on demand. Claude only reads the full skill content when it's relevant, so they don't waste your context window.

Where to find them:

Plugins worth trying:

  • Firecrawl - Web scraping and data extraction.
  • Ralph Loop - Autonomous coding loops. Claude iterates until specs are met.
  • Context7 - Feeds Claude up-to-date API docs so it's not working from stale training data.
  • Playwright - Browser testing automation from the terminal.

Key concepts to know:

  • CLAUDE.md is the most important file in your project. It stores your architecture, code standards, commands, and testing instructions. Claude reads it at the start of every session. Run /init to auto-generate one from your codebase.
  • Plan Mode (Shift+Tab twice) puts Claude in architect mode. It can research, analyze, and plan, but won't make changes until you approve.
  • Sub-agents are specialized Claude instances for specific tasks like debugging or security review. They get their own context so they don't pollute your main session.
  • Progressive disclosure means skills only use about 100 tokens at startup. The full content loads when Claude decides the skill is relevant.

Curated Community Lists

The community has built great resource collections. These are the ones worth bookmarking:

  • awesome-claude-code - 21.6k stars. The main curated list. Slash commands, CLAUDE.md files, CLI tools, workflows. Bookmark this one first.
  • awesome-claude - High-level view of the full Claude ecosystem. Models, interfaces, extensions, courses, community.
  • awesome-claude-code - Tools, IDE integrations, and frameworks. Organized with star count indicators.
  • awesome-claude-skills - 5.5k stars. Skills resources with a timeline of how the ecosystem evolved.
  • awesome-claude-skills - MCP Builder, Playwright, prompt engineering patterns, and domain-specific skills.
  • awesome-mcp-servers - 80.5k stars. The MCP servers list. Start here if you're connecting Claude to external tools.
  • awesome-claude-prompts - 3.8k stars. Prompt patterns you can adapt.

Community Guides & Best Practices

A lot of the best Claude Code content comes from people using it in production, not from Anthropic:

Community

Pricing

  • Free - Basic Claude Chat access with usage limits.
  • Pro ($20/month) - More usage, all models including Opus, skills, priority access.
  • Max ($100/month) - Maximum usage for heavy daily use.
  • Team - Shared workspaces, collaboration features, admin controls.
  • Enterprise - Full security, compliance, SSO, org controls. Nonprofit perks including free Opus 4.6 access.
  • API - Token-based. Opus at $5/$25 per million input/output. Sonnet and Haiku cheaper. Prompt caching available.

Current pricing at claude.ai/upgrade for plans and model docs for API rates.

← Back to Resource Directory