Claude Code
Published Updated
What is Claude Code?
Claude Code is Anthropic’s agentic coding tool: it reads your codebase, edits files and runs commands to complete multi-step engineering tasks. It runs in the terminal, as a desktop and web app, and inside VS Code and JetBrains.
An autonomous agentic coding assistant built by Anthropic that operates directly within the terminal to read codebases, edit files, and execute commands.
Claude Code started as a command-line interface (CLI) tool and now also ships as a desktop app (Mac/Windows), a web app, and IDE extensions (VS Code, JetBrains). Across these surfaces it functions as an autonomous agent within a developer's environment, navigating file systems, running tests, and utilizing native tools (like Git or grep) to complete complex software engineering tasks. It supports hooks, MCP servers, subagents, and skills for extending its behavior, and it requires explicit user permission before executing potentially destructive commands, maintaining a human-in-the-loop safety boundary.
How it relates to Orbit
Orbit can register Claude Code as a coding harness and show its live session in the project CLI tab. See Agent Sessions & Harnesses for details.
When you would use it
You use Claude Code when you want an AI to autonomously execute a multi-step engineering task directly from your terminal, rather than just auto-completing text in an editor.
Common operations
- Executing multi-file refactors across an entire repository.
- Piping error logs directly into the agent for automated debugging.
- Managing Git workflows via natural language.
Frequently asked questions
Is Claude Code free?
It is included with paid Claude plans such as Pro and Max, or billed per token through an Anthropic API key.
Can Orbit run Claude Code on my Claude subscription?
Yes. Orbit can drive Claude Code through your Claude Max or Pro plan using OAuth, so the work counts against the plan you already pay for. See Connect a Coding Agent.
How does Orbit use Claude Code?
Orbit registers Claude Code as a coding harness, routes complexity tiers to it, launches it for each approved prompt and shows the live session. See Agent Sessions & Harnesses.