Is Claude Code Free? Plans, Limits and API Charges

If you need the repository agent, the standard individual route is a paid Pro or Max subscription. An organization seat can include it too. The other route is a Claude Console API account, where usage is billed separately rather than drawn from the Free chat plan. Anthropic's current plan comparison marks Claude Code “No” under Free and “Yes” under Pro and Max.

Claude Free versus Claude Code: what is actually free?

Anthropic's pricing page lists free access to Claude on the web, desktop and mobile, including the ability to create files and run code. Its file-creation guide confirms code execution in chats is available to Free users. You can ask Claude to explain a Python error, draft a function, analyze an uploaded file or create an artifact without buying Pro, subject to the Free plan's usage allowance.

Claude Code is the separate coding agent that connects to a local or cloud repository, reads project files, proposes edits and works from a terminal or supported IDE. Anthropic's Pro and Max access guide describes that product and its subscription login. Free chat code execution does not turn a Free account into a Claude Code subscription. A free download of the Claude Code application likewise does not make the model usage free.

The distinction matters before you install anything. If your task is “explain this snippet” or “generate a small script I will inspect,” Free chat may be enough. If your task is “inspect this repository, change these files, run the project checks and show me the diff,” you are choosing a repository agent with different access and billing terms. Anthropic's plan page is the current source of truth for that access decision.

The official access routes, checked 20 October 2026

The prices below are the public US figures on Anthropic's pricing page, checked 20 October 2026. Taxes and local prices may differ. The API route has no flat subscription price in this table because the cost depends on the model and tokens used.

For an individual who only wants to try repository-level coding, Pro is the lowest standard listed subscription that includes Claude Code. The Pro plan guide confirms both Claude Code access and that a Pro subscription does not include Claude Console API usage. The annual option is $200 paid up front, which Anthropic displays as about $17 per month; monthly billing is $20. It is not a monthly $17 plan.

Max is a capacity decision, not a prerequisite for Claude Code. Anthropic lists Max from $100 per month and says it provides 5x or 20x the Pro session allowance, depending on the tier. The Max plan guide also lists Claude Code access. Someone whose coding work fits within Pro has no access reason to buy Max merely to unlock the agent.

Organization billing needs its own check. Anthropic's Team and Enterprise guide says every current Team seat includes Claude Code. New and self-serve Enterprise seats include it, while older Enterprise arrangements have different seat types. Ask the administrator which seat you have and whether the organization uses included usage, usage credits or usage-based billing. An old pricing screenshot is not enough to resolve a specific contract.

What happens when you reach a usage limit?

Paying for Pro does not buy unlimited agent time. Anthropic says a paid plan has a rolling five-hour session limit and a weekly limit, shared across Claude chat, desktop and Claude Code. Longer conversations, larger files, different models and more complex tasks consume that allowance differently. There is no dependable “X prompts per day” number that applies to every user or repository. Anthropic explains the distinction between usage limits and context length.

When a session allowance is exhausted, a reader has three normal choices: wait for the reset, use a higher-capacity eligible plan, or enable usage credits on a paid plan. The last option adds consumption-based charges. Anthropic's usage-credit guide says extra usage is charged separately and can be controlled with a monthly spend limit. Check your own Settings > Usage page for the live status rather than estimating from someone else's prompt count.

There is a second, separate paid path: authenticate Claude Code to a Claude Console API account. The Claude Code metering guide distinguishes subscription usage from API-key use, which is billed per token to the relevant account. Anthropic's API billing guide describes Console billing. API rates vary by model and features; read the current official API price table before setting a budget. Do not infer that a $20 Pro subscription also buys $20 of API usage.

The billing check that prevents an unpleasant surprise

An active Pro or Max subscription does not by itself prove that a particular terminal session is consuming subscription allowance. Anthropic warns that if an ANTHROPIC_API_KEY environment variable is present, Claude Code uses that API key instead of the subscription login. The result can be separate API charges. This is stated in its Pro and Max connection guide.

Before a long session, check the account Claude Code says it is using. If you intended subscription access but see Console billing, stop and review the login and environment configuration. Anthropic's guide says /login can switch from Console pay-as-you-go access to a subscription, and it documents /logout and reauthentication if the expected option does not appear. Do not paste keys into a chat or publish a screenshot that exposes them. For a subscription session, monitor the remaining allowance in Claude; for an API-key session, monitor the Console's cost and spend limit. Those meters answer different questions.

What the “free Claude Code” videos actually do

Some videos use the Claude Code interface through a third-party proxy or connect it to another model. That does not give you Anthropic's models for free or change Anthropic's plan entitlements. Check the service's terms before using it, and never paste API keys or account credentials into an unknown proxy. A trial or special offer may exist for an eligible account at a particular time; verify it on Anthropic's current page and read its end date and billing terms. Do not buy a shared login or assume an older video describes today's plan.

Which route should you choose?

Start with the task, not the plan name. For occasional code questions or a small script you will run yourself, try the free Claude chat and inspect the output. The article Can Claude Write Python Code? shows the distinction between generating code and checking it. If you want an agent to read and edit a repository repeatedly, compare Pro's included access against your expected interruptions. Look at your own usage after several real sessions before moving to Max.

If you need Claude Code for a short, irregular coding sprint, separately billed API access may suit you better than a subscription, but it is not automatically cheaper. Without a measured workload, any break-even claim would be guesswork. Set a Console spend limit, record one actual session's cost, and compare it with the published subscription terms. Anthropic's rate and spend-limit documentation explains that API spend caps and request rate limits are different controls.

For a team, confirm the seat and billing route before standardizing a workflow. Document who can enable extra usage credits, whether an API key is present on developer machines, where charges appear and who reviews the resulting code. A paid agent can reduce the time to a draft, but the team still owns the final diff and its tests. If you are learning the terminal workflow, the Claude Code QA guide demonstrates planning one small test and reading its failures rather than assuming the first generated test is correct.

The cost decision also has a publication consequence for teams creating coding guides. A new guide does not become useful merely because an agent wrote it. Check that the page answers a specific reader question, then monitor whether later AI answers cite its exact URL. Xtrusio's citation-grade content guide describes the source and claim check, while its AI visibility guide explains what to measure after publication. Neither check requires claiming that a coding tool caused a citation.

Sources and date

Prices and access were checked 20 October 2026 against Anthropic's pricing comparison, Pro and Max plan guides, Team and Enterprise guide, Claude Code subscription guide, usage-limit and credit guides, Claude Code metering guide, Console billing guide, API price table and API rate-limit documentation. These official pages are linked beside the relevant claims above. Plans and allowances can change; use the current provider pages when making a purchase decision.