# Claude Code — Agentic coding tool from Anthropic that edits your codebase, runs commands, and opens pull requests. > Source: The Agents Index — https://theagentsindex.com/claude-code (structured, researched, re-verified) > Facts last verified: 2026-09-02 Claude Code is Anthropic's agentic coding tool for developers. It reads a codebase, plans a change, edits files across it, runs commands and tests, and takes the work through to a commit and pull request. | Fact | Value | | --- | --- | | Website | https://claude.com/product/claude-code | | Price | from $20/month billed monthly, $17/month billed annually | | Free tier | No | Anthropic's $0 Claude plan does not include Claude Code: the pricing FAQ states 'Claude Code is included in all paid plans.' The cheapest way in is Pro at $17/month billed annually ($20 monthly), or pay-as-you-go API credits through a Console account. ## What it does **Writes code** · Yes — Takes a plain-language description and writes the change across multiple files, then verifies it. > Describe what you want in plain language. Claude Code plans the approach, writes the code across multiple files, and verifies it works. > — https://code.claude.com/docs/en/overview · What you can do, 'Build features and fix bugs' section **Debugs and fixes** · Yes — Takes an error message or symptom, traces it through the repository, and implements the fix. > Claude Code traces the issue through your codebase, identifies the root cause, and implements a fix. > — https://code.claude.com/docs/en/overview · What you can do, 'Build features and fix bugs' section **Writes tests** · Yes — Writes tests for untested code as a named use case, and runs them as part of the same task. > writing tests for untested code, fixing lint errors across a project, resolving merge conflicts, updating dependencies, and writing release notes > — https://code.claude.com/docs/en/overview · What you can do, 'Automate the work you keep putting off' section **Reviews code** · Yes — The /code-review command reviews a local diff on any plan; the managed GitHub PR reviewer is a research preview on Team and Enterprise. > Code Review analyzes your GitHub pull requests and posts findings as inline comments on the lines of code where it found issues. > — https://code.claude.com/docs/en/code-review · Intro paragraph, above 'How reviews work' **Whole-codebase work** · Yes — Dynamic workflows run many subagents over a repository for audits and large migrations; the vendor's own example is a 500-file migration. > A job outgrows a handful of subagents, or you want findings verified against each other: a codebase-wide audit, a 500-file migration, cross-checked research > — https://code.claude.com/docs/en/agents · Comparison table, Dynamic workflows row, 'Use it when' column **Ships it** · Partly — The claimed workflow ends at a commit, branch and pull request; deployment is only whatever shell command Claude is allowed to run. > integrates with GitHub, GitLab, and your command line tools to handle the entire workflow—reading issues, writing code, running tests, and submitting PRs—all from your terminal > — https://claude.com/product/claude-code · 'What could you do with Claude Code?' section, 'Turn issues into PRs' card ## How much it does unattended **Runs autonomously** · Yes — Auto mode has a classifier approve tool calls instead of a person; routines and background sessions run unattended, including in the cloud. > Auto-approves tool calls with background safety checks that verify actions align with your request > — https://code.claude.com/docs/en/permissions · Permission modes table, 'auto' row **Multi-agent** · Yes — Four coordination styles: subagents, background agent view, agent teams with inter-agent messaging (experimental, off by default), and dynamic workflows. > Subagents, agent view, agent teams, and dynamic workflows each parallelize work in a different way. > — https://code.claude.com/docs/en/agents · Opening paragraph **Agent permissions** · Yes — Starts read-only and asks per action in Manual mode; allow/ask/deny rules, six permission modes, and a bash sandbox set the boundary. > In Manual mode, Claude Code starts with read-only permissions. When Claude Code needs to edit files, run tests, or execute commands, it asks you first > — https://code.claude.com/docs/en/security · How we approach security, 'Permission-based architecture' ## Which models it runs on **Claude** · Yes — Claude is the only model family: Fable, Opus, Sonnet and Haiku aliases, resolved per provider. > Uses the latest Opus model for complex reasoning tasks > — https://code.claude.com/docs/en/model-config · Model aliases table, opus row **GPT** · No — On the first-party API the recognised model strings are aliases and names beginning claude-; the alias table lists only Claude families. > any name that starts with claude- > — https://code.claude.com/docs/en/model-config · Setting your model, list of model strings Claude Code recognizes on the Anthropic API **Open models** · No — No open-weight option: every documented provider path (Anthropic API, Bedrock, Google Cloud, Microsoft Foundry) serves Claude models only. > any name that starts with claude- > — https://code.claude.com/docs/en/model-config · Setting your model, list of model strings Claude Code recognizes on the Anthropic API **Model choice** · Partly — You choose the model and effort per session or per subagent, but only within Claude; admins can pin the allowlist. > Enterprise administrators can use availableModels in managed or policy settings to restrict which models users can select > — https://code.claude.com/docs/en/model-config · Restrict model selection, opening sentence ## Where you use it **In your editor** · Yes — First-party extensions for VS Code (and Cursor) and JetBrains IDEs, with inline diffs and plan review. > Native extensions for VS Code (+ Cursor, Devin Desktop) and JetBrains IDEs. > — https://claude.com/product/claude-code · 'Meets you where you code' section, 'Integrate with your editor' card **On the command line** · Yes — The terminal CLI is the primary surface, installed by shell script, Homebrew, WinGet or Linux package managers. > Super powerful terminal integration. Works with all your CLI tools alongside any IDE. > — https://claude.com/product/claude-code · 'Meets you where you code' section, 'Start in your terminal' card **In your pipeline** · Yes — GitHub Actions and GitLab CI/CD run it in your own pipelines; a managed GitHub PR reviewer exists on Team and Enterprise. > In CI, you can automate code review and issue triage with GitHub Actions or GitLab CI/CD. > — https://code.claude.com/docs/en/overview · What you can do, 'Create commits and pull requests' section **In a browser** · Yes — Cloud sessions run from claude.ai/code in a desktop browser or the iOS and Android app, with no local install. > Run Claude Code in your browser with no local setup. > — https://code.claude.com/docs/en/overview · Get started, Web tab ## Whose machine it runs on **Self-hosted** · Partly — The CLI executes on your own machine, and cloud sessions can run on your runners in public beta on Team and Enterprise; inference always goes to Anthropic or a cloud provider. > Self-hosted environments are in public beta on Team and Enterprise plans and are off by default. > — https://code.claude.com/docs/en/self-hosted-environments · Note at top of page **Open source** · No — The vendor's own repository carries an all-rights-reserved licence; the public repo holds plugins, docs and the issue tracker, not the CLI source. > © Anthropic PBC. All rights reserved. Use is subject to Anthropic's Commercial Terms of Service. > — https://github.com/anthropics/claude-code/blob/main/LICENSE.md · LICENSE.md, full file text ## What it costs to run **How it meters** · Yes — Two meters: a per-seat usage allowance on subscription plans, or per-token billing on the Console and cloud providers. > On Teams and Enterprise plans, usage draws from each member's seat allowance. On the Console and on cloud providers, usage is billed per token to your organization. > — https://code.claude.com/docs/en/costs · Manage costs for your organization, opening paragraph **Free tier** · Partly — Self-serve with no sales call, but there is no free tier: the $0 Claude plan excludes Claude Code, so the entry point is a paid plan or API pay-as-you-go. > Claude Code is included in all paid plans. > — https://claude.com/pricing · FAQ, Plans and usage tab, 'Is Claude Code included in my plan?' **API access** · Yes — The Agent SDK exposes the same agent loop as a Python or TypeScript library, and claude -p with JSON output drives it from any language. > the same tools, agent loop, and context management that power Claude Code, programmable in Python and TypeScript > — https://code.claude.com/docs/en/agent-sdk/overview · Opening paragraph **MCP server** · Yes — claude mcp serve exposes Claude Code as a stdio MCP server other clients connect to; it is also an MCP client for external servers. > You can use Claude Code itself as an MCP server that other applications can connect to > — https://code.claude.com/docs/en/mcp · 'Use Claude Code as an MCP server' section **Bring your own key** · Yes — Bring your own Anthropic API key, or your own cloud account via Amazon Bedrock, Google Cloud's Agent Platform or Microsoft Foundry; apiKeyHelper supports vault-issued tokens. > if your organization uses Amazon Bedrock, Google Cloud's Agent Platform, or Microsoft Foundry, set the required environment variables before running claude > — https://code.claude.com/docs/en/authentication · Log in to Claude Code, account types list, 'Cloud providers' bullet ## Buying it for a team **A company can buy it** · Yes — Team is self-serve for 2 to 150 seats; Enterprise is available self-serve or sales-assisted with central billing. > For teams of 2 to 150 > — https://claude.com/pricing · Team & Enterprise tab, Team plan card, subtitle **Seat model** · Yes — Per seat with a floor of 2 and a ceiling of 150 on Team; Standard $20/seat/mo annual ($25 monthly), Premium $100 ($125), Enterprise $20/seat plus usage. > Per seat / month if billed annually. $25 if billed monthly. > — https://claude.com/pricing · Team & Enterprise tab, Team plan card, Standard seat price note **Pooled budget** · No — Each member gets their own allowance on a rolling five-hour and weekly window; only usage-credit spend past it can be capped at org, group or member level. > each member's Claude Code usage draws from a per-seat allowance that resets on a rolling five-hour window and a weekly window > — https://code.claude.com/docs/en/costs · Claude for Teams and Enterprise section, first paragraph **Admin controls** · Yes — Managed and server-managed settings pin permission rules, model allowlists, telemetry and login method, and local settings cannot override them. > Security teams can configure managed permissions for what Claude Code is and is not allowed to do, which cannot be overwritten by local configuration. > — https://code.claude.com/docs/en/third-party-integrations · Best practices for organizations, 'Configure security policies' **Audit log** · Partly — Audit logs are an Enterprise plan feature, alongside a compliance API; cloud sessions log every operation, and OpenTelemetry export works on any plan. > Audit logs > — https://claude.com/pricing · Team & Enterprise tab, Enterprise plan card, feature list item **Single sign-on** · Partly — SSO ships on the business plans, not on individual Pro or Max; the vendor names no protocol, and SCIM is Enterprise-only. > Single sign-on (SSO) > — https://claude.com/pricing · Team & Enterprise tab, Team plan card, feature list item ## What happens to your code **Opt out of training** · Yes — Commercial terms (Team, Enterprise, API, cloud providers) are no-training by default; Free, Pro and Max users toggle it in privacy settings. > Anthropic does not train generative models using code or prompts sent to Claude Code under commercial terms, unless the customer has chosen to provide their data to us for model improvement > — https://code.claude.com/docs/en/data-usage · Data policies, 'Data training policy', Commercial users paragraph **Data residency** · Partly — US-only inference is selectable on the Claude API and Claude Platform on AWS at a 1.1x rate; workspace storage geo is US only, and Bedrock or Google Cloud regions follow the endpoint. > Inference runs only in US-based infrastructure. > — https://platform.claude.com/docs/en/manage-claude/data-residency · Inference geo table, 'us' row **Getting out** · Yes — /export writes a readable transcript, transcripts sit locally as JSONL, and org analytics and spend reports export as CSV; prepaid usage credits mostly do not expire. > Run /export to open a menu that lets you copy the current conversation to your clipboard or save it as a plain-text file > — https://code.claude.com/docs/en/sessions · 'Export and locate session data' section, first sentence **Certifications** · Yes — SOC 2 Type 2 and ISO 27001 artefacts are published at the Trust Center; a HIPAA-ready offering and zero data retention are Enterprise options. > Learn more and access resources (SOC 2 Type 2 report, ISO 27001 certificate, etc.) at Anthropic Trust Center > — https://code.claude.com/docs/en/security · How we approach security, 'Security foundation' ## Platform record (as of 2026-08-30) Reported by the code-hosting platform, not by the vendor. | Fact | Value | | --- | --- | | Stars | 143445 | | Forks | 22939 | | Last push | 2026-08-28 | ## Provenance Every fact above comes from the vendor. Nothing here is independently corroborated yet. ## Coverage elsewhere Pages about this product that the vendor does not control. The label states how independent each one is. | Source | Kind | Read | | --- | --- | --- | | anthropic.com | Independent source | 2026-08-22 | | anthropic.com | Independent source | 2026-08-22 | | anthropic.com | Independent source | 2026-08-22 | ## What changed - 2026-09-02 — Capabilities · Commercial terms · Free tier · Price - 2026-08-31 — Capabilities · Commercial terms · Free tier · Price - 2026-08-22 — Alternatives · One-liner · Summary ## Correcting this record Anything marked "Not established" is a question we have not answered — it is not a No. If you represent Claude Code, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth. Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.