Skip to content
AppDuo
ReviewsComparisonsBest ofCollections
Source-backed analysis12

Review basis

The AppDuo team prepared this review by analysing product features, official documentation and user feedback to help you understand its strengths, limitations and suitability for your needs.

12 reference sources

Information updated Nov 10, 2026

How we review
Claude Code logo

Claude Code Review

Anthropic's coding agent across terminal, IDE, web and CI

7.4/10

By · Updated

  • macOS
  • Windows
  • Linux
  • Web
  • iOS
  • Android
  • API
Paid · $20.00/mo
Visit Claude Code
Compare side by side
On this page

Decide at a glance

Our verdict

A fit for developers and teams who want to hand multi-step tasks (bug fixes, refactors, pull requests) to an agent working inside the repo, with an explicit permission model. The trade-offs: cost and usage depend on a Claude plan or API pricing, and it runs only Anthropic models.

Best for: Terminal-comfortable developers who want an agent to read, edit and run commands in their project

Skip it if: Anyone who wants a free tier: the Claude Free plan does not include Claude Code

Pros

  • Available on many surfaces: terminal, VS Code, JetBrains, desktop app, web, mobile, Slack and CI
  • Auto mode (the default in terminal and VS Code) uses a classifier to block risky actions; Manual mode asks before editing files or running commands
  • Extensible through CLAUDE.md, skills, hooks, MCP and the Agent SDK; bash commands can run in a sandbox

Cons

  • Usage is capped per 5 hours and per week; Anthropic added weekly limits for Pro and Max on 28 August 2025
  • Heavier-usage tiers are expensive: Max at 100 USD and Ultra at 200 USD a month
All pros and cons · Who is it for?

Pros

  • Detailed documentation that spells out data policy, retention and how to turn off telemetry

Cons

  • On individual plans, data can be used for training while that setting is on (and is then kept 5 years); local session transcripts are plaintext

Best for

  • Teams automating PR review and issue handling through GitHub Actions, GitLab CI/CD or Slack
  • Companies that need commercial data terms and want to run it through Amazon Bedrock, Google Cloud or Microsoft Foundry

Skip it if

  • People who want to pick models from several vendors inside one tool

Experience & key features

What Claude Code is

Claude Code is Anthropic's coding agent. According to the official documentation, it reads your codebase, edits files, runs commands and connects to development tools. It launched on 24 February 2025 as a limited research preview alongside Claude 3.7 Sonnet and became generally available on 22 May 2025 with Claude 4. The product page says it runs locally and talks directly to model APIs, with no backend server or remote code index.

How it works and key features

You install it with a one-line command, Homebrew, WinGet or apt/dnf/apk, then run claude in a project folder. Beyond the terminal there is a VS Code extension (which also works in Cursor), a JetBrains plugin, a desktop app, a web version at claude.ai/code and the Claude iOS and Android apps. Interactive terminal and VS Code sessions start in Auto mode, where a classifier reviews actions and blocks those it judges unsafe. In Manual mode it starts read-only and asks before editing files or running commands; bash commands can run in a sandbox. You customise it with CLAUDE.md, skills, hooks and MCP servers, run several agents in parallel, schedule recurring tasks, or automate through GitHub Actions, GitLab CI/CD and Slack. The official GitHub repository has over 155,000 stars.

Pricing and limits

Per the official pricing page, Claude Code is included in Pro (20 USD a month, or 17 USD a month billed annually), Max (100 USD), Ultra (200 USD), Team and Enterprise; the Free plan does not include it. Max and Ultra add 100 and 200 USD a month in API credits respectively. Team starts at 20 USD per seat per month billed annually (25 USD monthly) with up to 500 USD a month in pooled API credits. Limits reset every 5 hours and weekly. On 28 July 2025 Anthropic announced weekly limits for Pro and Max, effective 28 August 2025, estimated to affect under 5% of subscribers. A Console account is billed at standard API pricing.

More analysis

Compared with alternatives

Claude Code is a terminal-first agent that also plugs into CI, unlike IDE-centred AI editors such as Cursor or Windsurf. OpenAI Codex and GitHub Copilot also offer coding agents, with different models and pricing. Points to weigh are plan-based usage caps and the fact that it runs only Anthropic models. This review is compiled from documentation and public sources, not from hands-on use.

Screenshots

Claude Code in the terminal, from the official product page
Claude Code in the terminal, from the official product page
Claude Code in VS Code, from the official product page
Claude Code in VS Code, from the official product page
Claude Code on the web, from the official product page
Claude Code on the web, from the official product page

Plans & pricing

Prices last verified Nov 10, 2026

Pro

$20/mo

$17/mo billed yearly

20 USD a month, or 17 USD a month billed annually (200 USD up front). Includes Claude Code; usage limits reset every 5 hours and every week.

Max

$100/mo

100 USD a month. 5x Pro usage per 5-hour window and per week, plus 100 USD a month in API credits for the Claude Platform.

Ultra

$200/mo

200 USD a month. No 5-hour limit for verified accounts; weekly usage is 10x Pro; includes 200 USD a month in API credits.

Team (Standard seat)

$25/mo/user

$20/mo billed yearly

25 USD per seat per month, or 20 USD billed annually. For teams of 2 to 150; includes Claude Code and up to 500 USD a month in pooled API credits.

Team (Premium seat)

$125/mo/user

$100/mo billed yearly

125 USD per seat per month, or 100 USD billed annually. 5x the usage of a Standard seat; can be mixed with Standard seats.

Enterprise

$20/mo/user

From 20 USD per seat per month billed annually, 20-seat minimum, plus usage charged at API rates. Admins can set user and organisation spend limits.

Free plan limits
The Free plan does not include Claude Code. Through a Console (API) account it is billed at standard API pricing.
Free trial
No free trial
Cancellation & refunds

The pricing page states no refund or cancellation terms; it only lists billing cycles: Pro monthly or annual, Max and Ultra monthly.

Privacy

According to the official documentation, Claude Code runs locally but sends prompts and model outputs over the network, encrypted with TLS 1.2+. On Free, Pro, Max and Ultra plans, Anthropic trains on your data only when that setting is on: data is kept 5 years if it is on, 30 days if off. For Team, Enterprise and API use, Anthropic says it does not train on code or prompts unless the customer opts in, for example through the Development Partner Program; standard retention is 30 days, and zero data retention (ZDR) can be enabled for qualifying Enterprise accounts. Session transcripts are cached locally in plaintext for 30 days by default. Metrics (which never include code, prompts or file paths) are on by default with the Claude API and can be disabled with DISABLE_TELEMETRY=1. Transcripts sent via /feedback, /bug or /share are retained for 5 years.

Official pricing page

Claude Code alternatives

OpenAI Codex logoVSClaude Code logoOpenAI Codex vs Claude Code
GitHub Copilot logo7.5/10

GitHub Copilot

An AI coding assistant that runs in most major IDEs and ties into GitHub

Freemium
Cursor logo7.4/10

Cursor

An AI-first code editor with agents that run locally or in the cloud

Freemium
All alternatives

Frequently asked questions

Is Claude Code free?

No. The pricing page says the Free plan does not include Claude Code. You need Pro, Max, Ultra, Team or Enterprise, or a Console account billed at API prices.

Does Claude Code train on my code?

On Free, Pro, Max and Ultra it depends on your privacy setting. For Team, Enterprise and API, Anthropic says it does not train on code or prompts unless the customer opts in, for example through the Development Partner Program.

Which operating systems does it support?

The terminal version runs on macOS, Linux and Windows (including WSL). The desktop app is available for macOS and Windows, with an Ubuntu/Debian version in beta. There is also a web version and the Claude iOS and Android apps.

Does it edit files without asking?

In Auto mode, the default for interactive terminal and VS Code sessions, a classifier reviews actions and blocks those it judges unsafe, so many changes go through without a prompt. In Manual mode it starts read-only and asks before editing files or running commands. Accept Edits mode auto-approves file edits inside the working directory.

Review evidence & sources12

Scores

Ease of use
8.0
Why this score

Installs with one command, Homebrew or WinGet; you log in and run claude in a project folder. The desktop app and web version need no CLI. The interface is English-only. It still assumes terminal familiarity and an understanding of the permission modes.

Features
8.8
Why this score

The docs list parallel agents, scheduled tasks, hooks, skills, MCP, GitHub Actions, GitLab CI/CD, Slack, Chrome and the Agent SDK. Broader than most peers, though limited to Anthropic models.

Value for money
7.0
Why this score

Pro at 20 USD a month includes Claude Code, but 5-hour and weekly caps push heavy users to Max (100 USD) or Ultra (200 USD), which include 100 and 200 USD a month in API credits respectively. The Free plan excludes it.

Performance
7.3
Why this score

Anthropic's launch post says that in its early testing Claude Code finished in one pass tasks that would normally take 45+ minutes of manual work; that is a vendor figure. No independent benchmark appeared in the sources opened. User ratings are mixed: the VS Code extension holds 3.7/5 (851 ratings) and the beta JetBrains plugin 2.3/5 (483). Usage caps also limit throughput.

Support
6.2
Why this score

Extensive documentation, free Claude Academy courses, a public GitHub issue tracker and a HackerOne security program. The pricing page names no direct support channel per plan. Claude's Trustpilot profile (covering all Claude products, not Claude Code alone) is 1.5/5 from 2,134 reviews, with frequent complaints about unexpected charges, account blocks and support requests left unanswered for days.

Privacy & security
6.8
Why this score

Positives: commercial plans are not used for training, ZDR exists, telemetry can be disabled, and Anthropic offers a SOC 2 Type 2 report and ISO 27001 certificate. Negatives: individual-plan data can be used for training when the setting is on (5-year retention), local transcripts are plaintext, and /feedback transcripts are kept 5 years.

What users say

Sources

Specs & compatibility
Developer
Anthropic
Headquarters
United States
Launched
2025
Category
DevelopmentAI

Capabilities

  • AI features
  • Public API
  • Integrations
  • Real-time collaboration
  • Mobile app

Platforms

  • macOS
  • Windows
  • Linux
  • Web
  • iOS
  • Android
  • API

Interface languages

  • English

Integrations

  • VS Code
  • Cursor
  • JetBrains
  • GitHub Actions
  • GitLab CI/CD
  • Slack
  • MCP servers
  • Chrome
  • Amazon Bedrock
  • Google Cloud Agent Platform
  • Microsoft Foundry

Support

  • Help center
  • Community forum
  • Email

Change history

  1. Limits.Anthropic announced new weekly limits for Pro and Max (overall and Opus 4-specific), effective 28 August 2025.
  2. Feature.Claude Code became generally available with Claude 4, with beta VS Code and JetBrains extensions, a GitHub integration and an SDK.
  3. Launch.Anthropic introduced Claude Code as a limited research preview alongside Claude 3.7 Sonnet.
About the author