Claude Code

Purchase and access update

Answer first: For a monthly-billed individual subscription in this captured source set, Claude Pro is US$20 per month and includes Claude Code. Anthropic also offers a discounted annual Pro subscription, but the captured page does not state its numeric annual total. API-key use is billed separately, and an existing ANTHROPIC_API_KEY can route CLI use to API billing instead of the subscription allowance.

Check the login route: Anthropic says an ANTHROPIC_API_KEY environment variable takes precedence over a Claude subscription and causes API usage charges. Moving to API-priced credits requires explicit consent, while Console auto-reload is a separate setting to inspect.

Permissions matter before output quality

The permissions documentation says reads inside the working directory and added directories do not prompt in the documented manual mode, while shell commands, file modifications, web fetches outside preapproved documentation domains and web searches do. Saved allow rules can persist for a repository, while a file-modification approval lasts only for the session. Approval behavior depends on configuration.

For a first evaluation, use an approved disposable repository, begin with read-only or plan behavior, inspect the proposed commands and diff, and refuse production credentials. Kingy has not run Claude Code, measured task success, reviewed generated code, tested prompt-injection resistance or verified a security outcome.

The account type changes the data rule

Anthropic’s Claude Code data page says Free, Pro and Max data can be used to improve models when the user’s data-improvement setting is on. Under commercial terms—Team, Enterprise, API and supported third-party platforms—Anthropic says it does not train generative models on code or prompts unless the customer opts in. These are vendor policy statements, not an independent privacy audit.

The same page says local Claude Code sends prompts and model outputs over the network and stores local session transcripts in plaintext for 30 days by default. It also says sessions started or most recently continued in Claude Desktop or Cowork are exempt from that local cleanup limit by default. Cloud execution clones the selected repository to an isolated VM and follows the account’s retention and usage policy. Review account type, surface, local transcript cleanup, provider path, feedback controls, repository scope and organization policy together.

What this update leaves unknown

• No fixed task, token or message allowance is asserted; usage depends on model, context, repository and concurrent work.
• No Max currency code is inferred from a dollar symbol.
• No region-wide availability, tax-inclusive total, negotiated Enterprise price or cloud-provider price is asserted.
• No benchmark, popularity, security certification, code-quality score or recommendation is asserted.

Update sources and dates

Anthropic’s retained support-page metadata identifies the Pro page as last modified June 10, 2026 and the Max page as last modified August 7, 2026. Those are modification dates, not claimed publication dates. Kingy retrieved all nine first-party pages on September 8, 2026 UTC. When normalized source text changes, review the affected price, access, billing, permission or data-use claim before preparing another revision.

Current setup boundary

Anthropic’s Claude Code overview documents terminal, IDE, desktop and browser surfaces. Most surfaces require a Claude subscription or Anthropic Console account; the terminal CLI, VS Code and JetBrains also support third-party providers.

Authentication documentation describes Claude.ai, Teams/Enterprise, Console and cloud-provider routes; exact plan prices and usage limits are not asserted here.