Claude Code

anthropics/claude-codeReviewed by AI

What it does

Claude Code is Anthropic's coding assistant for the terminal. You open a project folder, type what you want in plain language, and it reads the code, changes files, runs commands and handles git work. It also plugs into editors and can be mentioned on GitHub. It is made for programmers. One command installs it on Windows, macOS and Linux, and then you sign in with a paid Claude plan or an API key. This repository holds the installers, documentation and the bug tracker, not the source code.

How easy it is to install

B · installer plus one simple step · Windows

Windows B
macOS B
Linux B

One command on every platform: `winget install Anthropic.ClaudeCode` on Windows, `brew install --cask claude-code` or the install script on macOS and Linux, and each release also has a ready binary. The extra step that keeps it off A is signing in with a paid Claude account or an API key, and everything after that is typed in a terminal.

Numbers on GitHub

Stars: 154,079on Oct 31, 2026
Downloads: 1,574,823114,529 this week
Growth: +601stars in 5 days

Stars over time

Facts

Runs on: Windows, macOS, Linux
You need to install: Nothing extra
Disk space: 147 MB–440 MB · from the installer
Accounts or keys: Anthropic Claude (a paid plan login or an API key)
Cost: Free, API costs extra · The program is free to download, but it only works with a paid Claude plan or Anthropic API credits. Use is governed by Anthropic's Commercial Terms of Service.
License: No license (No license)
Can replace: Copilot, Cursor
Last update: Oct 31, 2026

Risk signals (8)

AttentionInstall command runs a script from the internet
Install command pipes a remote script into a shell: curl -fsSL https://claude.ai/install.sh | bash
automatic check
Attentionclosed source despite being on github
LICENSE.md in full: "© Anthropic PBC. All rights reserved. Use is subject to Anthropic's Commercial Terms of Service." The top-level folders are documentation, plugins, examples and scripts; the program itself ships as a compiled bundle in the releases (claude-win32-x64.zip, 147 MB), so the code cannot be read or rebuilt from this repository.
AI review
Attentionsends usage and conversation data
README, "Data collection, usage, and retention": "we collect feedback, which includes usage data (such as code acceptance or rejections), associated conversation data, and user feedback submitted via the /bug command". The README states limited retention, restricted access and a policy against using feedback for model training.
AI review
Attentionedits files and runs commands on your machine
The project describes itself as "an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows". It reads and changes files in the folder you start it in and runs commands there.
AI review
InfoNo license detected
No license detected by GitHub, the license file, the LICENSES folder or the winget manifest
automatic check
Infowindows install also pipes a remote script
The Windows instruction is `irm https://claude.ai/install.ps1 | iex`, which downloads a script and runs it immediately, the same pattern as the macOS and Linux install line. `winget install Anthropic.ClaudeCode` is offered as an alternative.
AI review
Infovery large issue backlog
16,173 open issues against 154,080 stars. Bugs are reported here and with the /bug command, so the queue is big; it says nothing on its own about how fast they are handled.
AI review
Infonpm install route deprecated
README note: "Installation via npm is deprecated. Use one of the recommended methods below." Older guides that tell you to run `npm install -g @anthropic-ai/claude-code` are out of date.
AI review

Evidence is shown in its original language.

What Radar could not check

What you will actually spend: usage depends on how much code it reads and writes, and the plan or API credits are billed by Anthropic.
What the compiled bundle does, since the source is not published here.
Exactly which parts of a conversation leave your machine, beyond what the README and the data-usage page say.
Whether it behaves the same on Windows as on macOS and Linux, which are the platforms the README recommends first.