Claude vs Claude Code: Claude is Anthropic’s chat assistant on the web, desktop and mobile, for writing, research and analysis. Claude Code is the agentic coding tool that works inside your repository, edits files and runs commands from the terminal, IDE, desktop Code tab or web. Paid Claude plans include both, and they draw from one usage pool.
Applies to: Claude apps and Claude Code 2.1.287, per Anthropic’s docs, pricing page and help center as of October 2, 2026. Last checked: October 2, 2026.
On this page
The short difference
Anthropic’s help center puts it in one line: “Use Claude for writing, research, analysis, and more—at work and at home. Use Claude Code for your terminal-based coding workflows” (Use Claude Code with your Pro or Max plan).
The Claude Code docs describe Claude Code as “an agentic coding tool that reads your codebase, edits files, runs commands, and integrates with your development tools.” The chat apps answer in a conversation; Claude Code acts on a project folder, with a permission system that asks before it edits files or runs commands.
Comparison table
| Aspect | Claude (chat apps) | Claude Code |
|---|---|---|
| What it is | Chat assistant on web, iOS, Android and Desktop (Pricing) | Agentic coding tool for your codebase (Overview) |
| Where it runs | claude.ai, the mobile apps, the Desktop app’s Chat tab | Terminal CLI, VS Code and forks such as Cursor, JetBrains, the Desktop app’s Code tab, claude.ai/code and the Claude mobile app, Slack |
| What it works on | Your messages, uploaded files, connected apps | A project folder or repository, local or cloned in a cloud session |
| Typical actions | Answers, writing, research, Artifacts, files created with code execution | Edits files across a project, runs tests and shell commands, makes commits, branches and pull requests |
| Control model | Conversation | Permission modes (default, acceptEdits, plan, auto, dontAsk, bypassPermissions), allow/deny rules, sandboxing (Permission modes) |
| Project instructions | Projects, memory, user preferences | CLAUDE.md, auto memory, .claude/settings.json |
| Extensions | Connectors, skills, desktop extensions | MCP servers, skills, subagents, hooks, plugins, output styles (Extend Claude Code) |
| Automation | “Hand off and schedule tasks” on paid plans (Pricing) | claude -p, Agent SDK, GitHub Actions, routines and scheduled tasks |
| Free plan | Yes | No (Pricing) |
| Pro and Max | Yes | Yes, same subscription |
| Team and Enterprise | Yes | Yes, per the seat rules below |
| Billing | Claude plan (Free, Pro, Max, Team, Enterprise) | Paid Claude plan, or API billing through a Claude Console account or a cloud provider |
What the Claude apps are for
The pricing page lists the chat feature set: chat on web, iOS, Android and Desktop, web search, Artifacts, file creation with code execution, memory, projects, connectors, skills, voice mode and Research (paid plans). The Desktop app has three tabs: “Chat for conversations, Cowork for Dispatch and longer agentic work, and Code for software development” (Desktop application).
Use the chat apps when the output is text, analysis or a single file you will copy somewhere: drafting, explaining a concept, reviewing a snippet you paste, planning, research.
What Claude Code is for
Claude Code works on a whole codebase. The overview lists jobs such as writing and running tests, fixing lint errors across a project, resolving merge conflicts, updating dependencies, tracing a bug from an error message to a fix, and creating commits and pull requests:
claude "write tests for the auth module, run them, and fix any failures"
It also scripts and automates. Pipe input into it, or run it non-interactively in CI:
git diff main --name-only | claude -p "review these changed files for security issues"
Install it with the native installer from the overview:
curl -fsSL https://claude.ai/install.sh | bash
Then run claude in a project and log in with the same account you use for Claude. A tour of the features is in Claude Code features.
Where the two overlap
- Same account and app. The Desktop app holds both Chat and Code tabs; the Code tab requires a paid subscription (Overview).
- Claude Code on the web and mobile. Cloud sessions run at claude.ai/code and in the Claude app for iOS and Android. Remote Control lets you continue a local Claude Code session from your phone or a browser.
- Slack. In Code + Chat routing mode,
@Claudedecides per message whether to open a Claude Code session or a chat (Claude Code in Slack). - Skills. Both support skills; Claude Code skills follow the Agent Skills open standard (Skills).
How they share a subscription
Pro and Max
Anthropic: “With Pro and Max plans, you now have access to both Claude on the web, desktop, and mobile apps and Claude Code in your terminal with one unified subscription” (help center). To connect Claude Code, log in with the same credentials you use for Claude; if you are logged in through a Console account, run /login to switch.
The plan also covers Claude Code in VS Code, Cursor and other VS Code forks, and JetBrains IDEs, and “IDE usage counts toward the same usage limits shared across Claude and Claude Code.”
Plan details: Claude Code Pro, Max 5x, Max 20x. The Free plan does not include Claude Code.
Team and Enterprise
“Claude Code is included with every Team plan seat. Premium seats offer more usage for team members with heavier workloads” (Use Claude Code with your Team or Enterprise plan). On new and self-serve Enterprise plans, the single Enterprise seat includes Claude Code; older Enterprise plans include it on Chat + Claude Code seats or Premium seats. See Team Standard, Team Premium and Enterprise.
One usage pool
The usage limits article states: “your usage of all different Claude product surfaces (claude.ai, Claude Code, Claude Desktop) counts towards the same usage limit” (How do usage and length limits work?). The pricing page FAQ says the same: “your work in the terminal and your chats draw from one pool.” Limits reset on a rolling five-hour session window, and paid plans add weekly limits (Pricing). Check where you stand in Settings > Usage, or with /status inside Claude Code.
When you reach a limit on Pro or Max, the help center lists four options: upgrade, enable usage credits, switch to pay-as-you-go through a Claude Console account, or wait for the reset.
Warning
“If you have an ANTHROPIC_API_KEY environment variable set on your system, Claude Code will use this API key for authentication instead of your Claude subscription (Pro, Max, Team, or Enterprise plans), resulting in API usage charges rather than using your subscription’s included usage.” (Anthropic Help Center)
To keep Claude Code on your plan only, the help center says to decline the API credit option when offered, or log out and log back in with only your Pro or Max credentials:
claude logout
claude login
Claude Code without a subscription
Claude Code also runs on API billing through a Claude Console account (claude auth login --console), or through Amazon Bedrock, Google Cloud’s Agent Platform or Microsoft Foundry (CLI reference, Team setup). That usage is billed per token; see Claude Code with API billing and Anthropic’s pricing page for rates.
When to use which
| Job | Use | Why |
|---|---|---|
| Draft a document, research a topic, explain a concept | Claude | “writing, research, analysis, and more” |
| Ask about a snippet you paste | Claude | No repository access needed |
| Change code across several files and run the tests | Claude Code | Reads the codebase, edits files, runs commands |
| Fix a bug from a stack trace in your repo | Claude Code | Traces the issue through the codebase and implements a fix |
| Commit, branch and open a pull request | Claude Code | Works directly with git |
| Review every PR or triage issues in CI | Claude Code | GitHub Actions, GitLab CI/CD, claude -p |
| Run a coding task while away from your desk | Claude Code on the web or mobile | Cloud sessions continue after you close the app |
| Long agentic non-coding work on the desktop | Claude Desktop, Cowork tab | Listed as the tab for “longer agentic work” |
