Anthropic Claude Code
Also known as: Claude Code
Claude Code is Anthropic's agentic coding tool. It reads a whole codebase, plans changes, edits files across the project, runs tests and commands and commits the result, working from the terminal, VS Code and JetBrains IDEs, the Claude desktop app, the web, mobile and Slack. Teams extend it with CLAUDE.md instructions, skills, subagents, hooks and plugins, connect it to their systems over MCP, fan large jobs out to hundreds of subagents and schedule cloud tasks, under permission modes from Manual approval to a classifier reviewed auto mode.
Claude Code is the agentic coding tool from Anthropic, the San Francisco company that makes the Claude models. A developer describes a task in plain language and Claude Code carries it out across a real codebase: it reads the project, plans an approach, edits files across it, runs tests and commands, fixes failures and commits the result.
It runs in the terminal, in VS Code and JetBrains IDEs, in the Claude desktop app, on the web, on iOS and Android, in Slack and in GitHub. Ramp uses it to turn notebook analysis into Metaflow pipelines and saves one to two days of routine work per model, and at Notion engineers decide what needs doing while Claude Code builds and verifies the work end to end.
Interactive terminal and VS Code sessions start in auto mode, where a separate classifier model reviews each action before it runs and blocks what it judges unsafe, and an organization can turn auto mode off in managed settings. Manual mode asks before edits, shell commands and network requests, Plan mode proposes steps without changing anything, and checkpoints and sandboxing add further control. CLAUDE.md files hold a repository's standing conventions, auto memory keeps notes Claude writes as it works, and teams add skills, subagents, hooks and slash commands, bundled as plugins from Anthropic's official marketplace.
MCP connects it to databases, internal APIs, issue trackers and monitoring systems, and a Chrome integration lets it drive a real browser. For large jobs, dynamic workflows run dozens to hundreds of subagents from a script Claude writes, and cloud sessions run in isolated Anthropic managed VMs or on the organization's own infrastructure, with cloud scheduled tasks starting work on a schedule, an HTTP call or a GitHub event. The Claude Agent SDK, a separate product, packages the same agent loop as a library for teams building agents of their own.
Claude Code comes with paid Claude plans rather than on its own. Pro is $17 a month billed annually or $20 monthly, Max is $100 a month with 5x Pro usage and $100 in API credits, Ultra is $200 a month with $200 in API credits, Team seats run $20 or $100 billed annually, and Enterprise is $20 a seat billed annually with a 20 seat minimum plus usage at API rates; the Free plan does not include it. It can also bill per token on the Claude API, Amazon Bedrock, Google Cloud or Microsoft Foundry. Anthropic holds SOC 2 Type 2 and ISO 27001, does not train on commercial Claude Code data unless the customer opts in, and offers Zero Data Retention to qualified Enterprise accounts.
It fits engineers handling large refactors, debugging and code review, teams that want an agent in CI and on a schedule, and founders and product managers who describe an outcome and want working software back. It fits less well for teams that need models from other makers, a persistent search index over company documents, or flat spend on heavily automated API runs.
From our comparison video
Claude Code vs Codebuff: 7 Questions for Engineering Teams Picking a Coding Agent
Our 8 minute video puts Anthropic Claude Code side by side with Codebuff. It covers what each one does and how each is priced, then when to choose which.
Recorded in October 2026. The profile here is live and updates as research is verified.
Vendor details
Canonical URL
https://code.claude.com/docs/en/overview
Category
Coding agent
Subcategory
Anthropic's agentic coding tool for whole codebases, across the terminal, IDEs, desktop app, web and Slack, with subagents, plugins, MCP and scheduled cloud tasks
Funding status
Product of Anthropic PBC, the San Francisco public benefit corporation that makes the Claude models.
Company status
first party product
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
MCP client support reaches databases, internal APIs, issue trackers and monitoring systems, with project, user, local and managed configuration scopes, OAuth 2.0 authentication and enterprise allowlists and denylists. Connectors in the Anthropic Directory are reviewed against published listing criteria, though Anthropic does not audit the security of MCP servers. Built in file, search and Bash tools, LSP servers, and a Chrome integration for driving a real browser session come with it. Claude Code runs in the terminal, desktop, IDEs, web, GitHub Actions and via Remote Control, and against Claude subscriptions, the Claude API, Amazon Bedrock, Google Cloud's Agent Platform or Microsoft Foundry.
In practice
A failing test comes from an issue spread across several files, and you don't want to trace it by hand. Claude Code reads the whole project, plans a fix, edits the relevant files, runs the suite, and iterates until it passes.
You won't let an agent touch production code unsupervised. In Manual mode Claude Code asks before it changes files or runs commands, and Plan Mode lets you review its proposed steps before any of them run.
Before writing a migration, the agent needs to know your actual data model and the ticket's acceptance criteria. Through MCP, Claude Code reads your database and issue tracker, not just the code in front of it.
Sources & related URLs
Research sources
Agentic Index coverage score
12.5 / 14 capabilities · 89%
| Integrations & Tool Calling | Full |
|---|---|
|
MCP servers connect the agent to tools, databases, APIs, issue trackers and monitoring systems at local, project and user scopes, with OAuth 2.0 sign in for remote servers, claude.ai connectors, and managed settings that allow or deny servers across an organization, alongside built in file, search, shell and web tools. Sourcecode.claude.com/docs/en/mcp and docs/en/chromeread 2026-10-24 |
|
| Workflow Orchestration | Full |
|
Subagents are workers Claude spawns for delegated tasks, and agent teams let a lead agent supervise long running peer sessions that share a task list. Dynamic workflows go further. Claude writes a JavaScript script that runs dozens to hundreds of subagents in the background, up to 1,000 per run and 16 at a time by default, asks before a run starts (every run in Manual mode, the first launch in Auto mode), and the script can be saved and rerun as a command. Sourcecode.claude.com/docs/en/workflowsread 2026-10-24 |
|
| Knowledge Grounding & RAG | Partial |
|
The agent grounds itself in the customer's repository by reading and searching it directly on each run, with project instructions from CLAUDE.md or AGENTS.md and path scoped rules, and MCP connections reach the customer's databases, issue trackers and internal APIs. Context is gathered fresh on each run, with no index, graph or embeddings layer kept between runs. Sourcecode.claude.com/docs/en/memory and docs/en/mcpread 2026-10-24 |
|
| Human Oversight & Guardrails | Full |
|
Auto mode is the starting mode for interactive terminal and VS Code sessions from v2.1.283. A separate classifier model reviews each action before it runs and blocks what it judges unsafe, the user's ask and deny rules still apply, and an organization can turn auto mode off in managed settings. Manual mode starts read only and asks before edits, shell commands and network requests, with approve once or allow from then on and fail closed matching for unmatched commands. Plan mode researches and proposes changes without editing, and claude -p and Agent SDK runs start in Manual. Sourcecode.claude.com/docs/en/permission-modes and docs/en/securityread 2026-10-24 |
|
| Security, Identity & Governance | Full |
|
A SOC 2 Type 2 report and an ISO 27001 certificate sit in the Anthropic Trust Center. Claude Code adds permission modes, sandboxed shell commands with filesystem and network isolation, command injection detection, trust verification for new folders and MCP servers, keychain credential storage, managed settings for organization enforcement and a HackerOne disclosure program. Under commercial terms Anthropic does not train models on Claude Code code or prompts unless the customer opts in. Sourcecode.claude.com/docs/en/security and docs/en/data-usageread 2026-10-24 |
|
| Observability & Auditability | Full |
|
OpenTelemetry exports metrics, events and optional traces, with a tool decision event for every permission accept or reject and a tool result event for every completed tool call, and OTEL_LOG_TOOL_DETAILS adds commands, MCP server and tool names and inputs. Telemetry endpoints set in managed settings override developer settings, and all operations in cloud sessions are logged for audit. Sourcecode.claude.com/docs/en/monitoring-usage and docs/en/securityread 2026-10-24 |
|
| Memory & State Persistence | Full |
|
Auto memory, on by default in local sessions, holds notes Claude writes itself as it works, such as corrections the user gives and approaches the user confirms. It is scoped per repository and shared across worktrees, loaded into every session through a MEMORY.md index, kept outside the transcript retention sweep until the user or Claude edits or deletes it, and browsable and editable through /memory. CLAUDE.md files are standing instructions the user writes, separate from this memory. Sourcecode.claude.com/docs/en/memoryread 2026-10-24 |
|
| Deployment & Data Residency | Full |
|
Claude Code runs on the developer's own machine by default, and Remote Control keeps code execution and file access local. Cloud sessions run in isolated Anthropic managed VMs with network access limited by default, or an organization routes them, scheduled tasks included, to self hosted environments on its own infrastructure. Prompts and outputs travel over TLS 1.2 or later, and at rest the Anthropic API uses AES-256 disk encryption, while Amazon Bedrock, Google Cloud's Agent Platform and Microsoft Foundry apply their own keys, with customer managed keys available on the first two. Commercial data is kept 30 days by default, and Zero Data Retention is turned on per organization for qualified Claude for Enterprise accounts. Sourcecode.claude.com/docs/en/data-usage and docs/en/securityread 2026-10-24 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
Plugins package skills, agents, hooks and MCP servers as one installable unit, and Anthropic's official marketplace is added the first time an interactive terminal session starts, browsable from the /plugin Discover tab or Claude Marketplace on the web, with a context cost estimate for each official plugin and named plugins such as commit-commands. Saved workflows and the bundled /deep-research workflow run as commands, and workflows can ship inside plugins. Sourcecode.claude.com/docs/en/plugins and docs/en/workflowsread 2026-10-24 |
|
| Triggers & Channel Coverage | Full |
|
Cloud scheduled tasks, in research preview, run a saved prompt against chosen repositories on a schedule, on an HTTP POST to the task's own endpoint, or on GitHub pull request and release events, so they keep working with the laptop closed. GitHub Actions and GitLab CI/CD automate code review and issue triage in CI, GitHub Code Review reviews every pull request automatically, and the agent answers @Claude in Slack and runs from the terminal, desktop, IDEs, the web and headless scripts. Sourcecode.claude.com/docs/en/routines and docs/en/overviewread 2026-10-24 |
|
| Model Flexibility & Routing | Not documented |
|
Every model Claude Code runs is Anthropic's. The customer picks among Claude models per session or subagent and picks where they are served (the Claude API, Amazon Bedrock, Google Cloud's Agent Platform or Microsoft Foundry), but not the model maker, and Anthropic does not support routing Claude Code to other makers' models through any gateway. Sourcecode.claude.com/docs/en/llm-gatewayread 2026-10-24 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
Claude Code runs programmatically from the CLI with -p, returning text, JSON with the session ID and cost, or newline delimited stream JSON, validating output against a JSON Schema, resuming sessions by ID, and taking permission modes, allowed tools and a permission prompt tool for unattended runs in scripts and CI. The Python and TypeScript packages of the Claude Agent SDK, a separate product, drive the same agent loop with hooks, subagents, MCP and session control, and each cloud scheduled task gets its own HTTP endpoint that starts a run. Sourcecode.claude.com/docs/en/headless and docs/en/routinesread 2026-11-18 |
|
| Testing, Debugging & Optimization | Full |
|
claude plugin eval runs a suite of realistic prompts against a plugin and scores the results, grading each case with a regex over the reply, a check that a particular tool was called, or a rubric a second model judges. It compares the run against a no plugin baseline and gates CI on the score, so a team can catch regressions when its plugin changes or a new model ships. Sourcecode.claude.com/docs/en/plugin-evalsread 2026-10-24 |
|
| Browser & Computer Use | Full |
|
The Chrome integration lets the agent work in a visible browser window with the user's logins, navigating, clicking, typing, filling forms, uploading files, reading console errors and DOM state and saving screenshots, in Chrome, Edge and other Chromium browsers. Site permissions come from the extension, and the agent pauses for logins and CAPTCHAs. Computer use extends this to native macOS apps. Sourcecode.claude.com/docs/en/chromeread 2026-10-24 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Product updates
Claude Code 2.1.325 changes sandboxed Bash commands to follow the normal permission flow by default, so they no longer run without a prompt unless sandbox.autoAllowBashIfSandboxed is set to true. The same release makes scheduled tasks and /loop ask before running a skill that declares allowed-tools or hooks, and makes read only git commands ask first in repositories outside the project.
Bears on: Human approval / guardrails
View sourceClaude Code 2.1.323 adds Code Mode, an experimental opt in in which Claude writes JavaScript that calls its tools instead of requesting each tool call one at a time. One piece of code can run commands in parallel, feed one result into the next call and return only the output Claude needs. It is turned on with CLAUDE_CODE_EXPERIMENTAL_CODE_MODE=1 in an interactive terminal session, and organizations can block it. The release also adds Claude Tag sessions started in Slack to the Compliance API on Enterprise plans, in beta.
Bears on: MCP / tool calling / API
View sourceA new permission setting in Claude Code 2.1.319 denies any tool call that would otherwise wait on a permission prompt and lets the turn carry on, in terminal sessions, subagents and teammates. The same release runs MCP tasks in the background: when an MCP server handles a tool call as a task, Claude Code moves it aside right away and tells Claude the result when it finishes.
Bears on: Human approval / guardrails
View sourcePricing
Pro $17 a month annually or $20 monthly; Max $100; Ultra $200; Team and Enterprise per seat
A subscription per user or per seat, within plan usage limits, or per token on the Claude API.
What is public
Anthropic sells Claude Code only as part of its paid Claude plans or on the Claude API, with no standalone price, and the Free plan does not include it. Pro costs $17 a month billed annually or $20 monthly. Max costs $100 a month and includes $100 a month in Claude Platform API credits. Ultra costs $200 a month, with no five hour limit and $200 a month in API credits. Team, for 2 to 150 people, has Standard seats at $20 and Premium seats at $100 a month billed annually, or $25 and $125 monthly. Enterprise is $20 per seat a month billed annually, with a 20 seat minimum, plus usage.
Billing mechanics
Usage limits on each plan reset in rolling five hour sessions, with weekly limits as well. Max carries 5x the usage of Pro, and Ultra carries 10x Pro's weekly usage with no five hour limit for verified accounts. Team plans pool up to $500 a month in API credits across the team. Enterprise usage on top of the seat price is billed at API rates that vary by model and task.
Cost watchouts
The API path has no hard cap, so automated or high volume agent runs can climb well past the subscription price.
Variable cost rationale
A subscription holds spend at the plan price. On the API and for Enterprise usage, spend grows with the number and size of agent runs and the model used.
Overage / add-ons
Past the limits, paid plans buy optional usage credits at Claude API rates, or move to Max or Ultra.
Sales call required
No, self serve available
Free / trial
There is no Claude Code trial.
Cancellation / refund
Billing runs through the Claude subscription or API account under Anthropic's standard terms, and Enterprise is billed annually only.
Missing data
Usage limits come as multiples of Pro rather than unit counts, and there is no fixed point where API metering begins.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose AI agents build, host and manage…
- AgentUI — Managed AI app builder for operations teams that generates and hosts…
- Aider — Open source, model agnostic terminal coding agent that edits your…
- Amp — Amp is a coding agent from Amp Frontier Corporation that runs each…
- Anything — Anything is an AI app builder whose agent turns a plain language…
Alternatives to Anthropic Claude Code
Other products that do the work Anthropic Claude Code does, and what each one is.
- LovableNatural-language full-stack app builderAnthropic Claude Code vs Lovable →
- OpenAI CodexAnthropic Claude Code vs OpenAI Codex →
- CognitionMaker of DevinAnthropic Claude Code vs Cognition →
- CursorAI native editor and agent platformAnthropic Claude Code vs Cursor →
- GitHub CopilotAnthropic Claude Code vs GitHub Copilot →
- ClineOpen source Apache 2.0 coding agent runtime across VS CodeAnthropic Claude Code vs Cline →