What it does
Describe a coding task in plain words and Claude Code works through it in your project: it reads the codebase, edits files, runs commands and tests, and can make commits and open pull requests.
Checked 2026-11-14 · Source · Source 2Good for you if…
- You already pay for Claude Pro, Max or Ultra: Claude Code is included
- You want an AI to take on a whole coding task in your project: reading the code, editing files, running tests
- You work in the terminal, VS Code or JetBrains, or want to hand off tasks from the web or your phone
Think twice if…
- You want to try it for free: the free Claude plan doesn’t include Claude Code
- You want unlimited use: Pro has usage limits that reset every 5 hours and every week
- Your code must never leave your computer: Claude Code sends your prompts, and the code it works on, to the model over the internet
How to start
- Get Claude Pro or a higher plan at claude.com/pricing (the free plan doesn’t include Claude Code), or set up a Claude Console account to pay per use.
- Install Claude Code by following Anthropic’s official install guide. Prefer not to use a terminal? Use the Code tab in the Claude desktop app, or the VS Code or JetBrains extension.
- Open a terminal in your project folder and type claude. It asks you to sign in to your Claude account the first time.
- Describe the task in plain words, for example “fix the login bug” or “write tests for this module”. Claude Code reads the code, plans, edits files and runs commands.
- Check the changes it made before you keep them.
Official install guide Checked 2026-11-14 · Source
Will it train AI on my code?
It depends on your account. With a personal Claude plan (Pro, Max or Ultra), Anthropic trains new models on your Claude Code sessions when the model-training setting is on. New users pick this when they sign up; Anthropic’s pages don’t say whether it starts on or off. You can turn it off any time under Settings › Privacy › “Help Improve our AI models”. Even when it’s off, sessions you send as feedback, or that are flagged for safety review, can still be used. With Team, Enterprise or an API (Console) key, Anthropic doesn’t train on your code or prompts unless your organization opts in (for example, the Development Partner Program).
Applies to Claude Code (terminal, IDE, desktop app or web) sending your code to Anthropic’s own Claude models. Personal Pro, Max and Ultra accounts follow Anthropic’s Consumer Terms; Team, Enterprise and API accounts follow its Commercial Terms. When you run Claude Code through Amazon Bedrock, Google Cloud or Microsoft Foundry, Anthropic says it doesn’t train on that data either; what those cloud providers do with it is not checked.
Anthropic does not train generative models using code or prompts sent to Claude Code under commercial terms, unless the customer has chosen to provide their data to us for model improvementIf you’re a new user, you can pick your setting for model training during the signup process.
Source 2Under Help Improve our AI models, click the button to toggle it off/on
Source 3Even if you opt out, we will use Materials for model training when: (1) you provide Feedback to us regarding any Materials, or (2) your Materials are flagged for safety review
Source 4Commercial users: (Team and Enterprise plans, API, 3rd-party platforms, and Claude Gov) maintain existing policies
Source 5Anthropic may not train models on Customer Content from Services.
Source 6Claude Code runs locally. To interact with the LLM, Claude Code sends data over the network.
Source 7
Summary of the maker’s terms as checked on 2026-11-14. Always read the official terms; this is not legal advice.
