Claude is the app you talk to on the web, desktop and phone: you ask questions, write, analyze files, and (since September 2026) hand it longer tasks in the same message box. Claude Code is Anthropic's coding agent: it works inside your code repository from the terminal, your IDE, the desktop app's Code tab or the web, editing files, running tests and committing. Same models, same subscription from Pro up; the free plan includes Claude but not Claude Code.
Claude and Claude Code at a glance
As of October 2026, from claude.com and Anthropic's docs.
| Claude (the app) | Claude Code | |
|---|---|---|
| What it is | Assistant for questions, writing, analysis and knowledge-work tasks | Agent for software work in a repository |
| Where you use it | claude.ai, desktop app (macOS, Windows; Linux in beta), iOS and Android | Terminal, VS Code, JetBrains, desktop app (Code tab), claude.ai/code, mobile app, Slack |
| Free plan | Yes | No |
| Paid plans | Pro $20 a month, Max from $100, Team, Enterprise | Same plans, same usage limits; or an API key |
| Works with | Uploads, connectors, folders you connect (Cowork), browser and computer use | Your repository, shell commands, git, MCP servers |
| Runs commands on your machine | No terminal; Cowork can use a browser and, in research preview, your computer's apps | Yes, in your project, with permission modes |
| Extensions | Connectors, skills, plugins | MCP servers, skills, plugins, hooks, subagents |
| Scheduling | Scheduled tasks (Cowork) | Routines in the cloud, desktop scheduled tasks, /loop |
| Output | Answers, documents, spreadsheets, slides | Commits, branches, pull requests |
What "Claude" means in 2026
The Claude app used to have two modes: chat, and Cowork for longer tasks. On September 16, 2026, Anthropic started merging them into "one Claude", rolling out to Pro and Max first. You type into one box and Claude decides whether a quick answer is enough or whether it should take on a task, work through it, and come back with a file. Tasks can keep going in the cloud after you close your laptop. The same update added Claude Docs and Claude Slides (in beta on paid plans) for editing and presenting what Claude makes.
So the Claude app is no longer "just chat". It can research, build a spreadsheet with working formulas, draft a deck, and act in a browser. What it is not built for is changing a codebase: there is no terminal workflow, no test loop and no git.
What Claude Code adds
Claude Code runs the same models with a developer's toolset:
- It reads the whole repository and
CLAUDE.md(your project rules) at the start of each session. - It edits many files in one task, runs your build and tests, and fixes what fails.
- It works with git: branches, commit messages, pull requests.
- It shows its edits as inline diffs in VS Code and JetBrains, and asks before risky steps according to the permission mode you pick.
- It connects to services through MCP servers and runs your hooks on 33 lifecycle events, so rules like "run the linter after every edit" always hold. See our Claude Code hooks guide.
- It runs unattended: cloud sessions on claude.ai/code, and routines that start on a schedule, an HTTP call, or a GitHub event.
A typical first run looks like this:
cd your-project claude "write tests for the auth module, run them, and fix any failures"
Claude desktop vs Claude Code
This is where most of the confusion comes from, because the Claude desktop app contains both. Open it and you get the regular Claude experience; click the Code tab and you are in Claude Code, with no separate install. The Code tab adds visual diff review, several sessions side by side, recurring scheduled tasks and cloud sessions. It needs a paid plan.
The terminal CLI is the same engine with the most options (every flag, scripting with claude -p, piping logs in). You can move between the two: run /desktop in a terminal session to continue it in the desktop app (macOS and x64 Windows), or pick up a session from your phone with Remote Control.
Same plan, shared limits
Claude Code doesn't cost extra on Pro, Max, Team or Enterprise, but it draws on the same usage. Pro gives at least five times the free plan's usage per five-hour session, Max five or twenty times Pro, with weekly limits on paid plans. Long agent sessions use more than chat, and so do Cowork tasks, so if you do a lot of both, Max gives you the most room. Developers who prefer pay-per-token can use Claude Code with an Anthropic Console API key instead.
Three tasks, three tools
- "Explain this stack trace." Either works. Paste it into Claude, or ask Claude Code inside the repo so it can open the files the trace mentions.
- "Turn these 30 customer interviews into a findings deck." The Claude app, as a task.
- "Add pagination to the orders API and open a PR." Claude Code.
Which to choose
- You want answers, writing help, or a document at the end: Claude.
- You want changes made in a codebase: Claude Code.
- You are on the free plan: Claude only; Claude Code starts at Pro.
- You like a GUI: use the desktop app's Code tab rather than the terminal.
- You want to compare the two agents inside the app: read Claude Cowork vs Claude Code, or Claude chat vs Cowork for the app itself.
Questions
Is Claude Code included in Claude Pro?
Yes. Claude Code is included in Pro ($20 a month), Max, Team and Enterprise, and shares the plan's usage limits. It is not in the free plan.
Do I need to install Claude Code if I have Claude desktop?
No. The Claude desktop app includes Claude Code in its Code tab, so you don't need the CLI separately. You need a paid plan to use it.
Is Claude Code a different model from Claude?
No. Both use the same Claude models (Opus 5.5 is the default for Claude Code on paid plans). The difference is the tools around the model: Claude Code can read your repository, edit files and run commands.
Can I use Claude Code without coding experience?
You can, but it is built for software work in a repository. For documents, spreadsheets and research tasks, the Claude app's Cowork mode is usually the better fit.