Claude Code
Anthropic's official agentic CLI for delegating coding tasks from the terminal.
From $20/mo
Features
- ▸ Subagents (custom roles)
- ▸ Slash commands
- ▸ Hooks for lifecycle events
- ▸ MCP server support
- ▸ Parallel agent execution
- ▸ Plan mode for review-before-apply
- ▸ Multi-file edits with diff approval
Pros
- + Tightest integration with frontier Anthropic models
- + Excellent at long, multi-step refactors
- + First-class subagent + plugin ecosystem
Cons
- − CLI-first; UI surface limited to terminal
- − Subscription-only, no community version
Best for
- · Senior engineers refactoring large codebases
- · Agentic workflows
- · Plugin authors
Avoid if
- · You need a free OSS alternative
- · You prefer point-and-click IDE UX
Recipes using Claude Code
Claude Code + Linear MCP — turn tickets into PRs in one slash
Wire Claude Code's MCP support to Linear so /feature pulls a ticket, implements, opens a PR, and links back.
When to reach for Cursor vs Claude Code (decision tree)
Both tools are great. They're great at different things. Here's how to pick in 30 seconds.
Build a multi-agent team in OpenCode (or Claude Code) for your project
Architect → Developer → Tester → Reviewer chain. Same pattern works in both.
Setting a sane budget for AI coding tools as a working engineer
Stop guessing — concrete monthly numbers for solo devs, small teams, and consultancies.