AI Agent Index
ByHeather MacAvelia·Independently reviewed·Published Mar 18, 2026·Updated Aug 24, 2026
Independently verified against live vendor data on Aug 12, 2026.

Anthropic's agentic coding agent for terminal, VS Code, JetBrains, Slack, web, desktop and iOS. Pro from $17/mo billed annually (USD), Max $100 to $200/mo. MCP server and client.

How we scored it

Autonomy

4/5

Integrations

5/5

Pricing clarity

4/5

Evidence

5/5

Setup

4/5

The facts

Claude Code is Anthropic's agentic coding tool. It runs in the terminal and also in VS Code, JetBrains IDEs, Slack, a web browser, the desktop app and on iOS. Unlike autocomplete-focused assistants, it reads the whole codebase, proposes multi-file changes, runs shell commands, edits files and handles git workflows through natural language. The developer reviews and approves changes by default, and an opt-in auto mode removes the per-step prompt for long-running work. The product is built on Anthropic's own Claude models, including Fable 5, Opus 5, Sonnet 5 and Haiku 4.5. Common workflows include refactoring across many files, debugging from a natural-language description, generating test coverage, writing and running scripts, and managing git operations including commits, branches and pull requests. Background agents let Claude Code work on a task while the developer does something else. It is an MCP client and an MCP server at the same time. It connects out to external MCP servers over HTTP, SSE, stdio and WebSocket transports, and the claude mcp serve command exposes Claude Code itself as a stdio server that other applications can connect to. GitHub Actions integration lets it run in CI without a local developer session. Pricing is in USD and follows the Claude subscription tiers rather than a standalone Claude Code plan. Pro is $17 per month billed annually at $200 up front, or $20 month to month. Max 5x is $100 per month and Max 20x is $200 per month, and Anthropic notes that both are web subscription rates which may differ on mobile app stores. Team standard seats are $20 per seat per month billed annually or $25 billed monthly, Team premium seats are $100 or $125 on the same basis, and Team plans require a minimum of two members. Enterprise is $20 per seat per month plus usage billed at API rates, billed annually, and can be started self-serve or through sales. Pay-as-you-go API access is priced per million tokens at $2 input and $10 output for Sonnet 5, $5 and $25 for Opus 5, $10 and $50 for Fable 5, and $1 and $5 for Haiku 4.5. What the entry price does not buy is a separate allowance. Claude Code has no standalone plan and shares one usage pool with everything else on the subscription, so terminal sessions and chats draw down the same rolling five-hour and weekly limits. There is no free tier of any kind: the Free plan excludes Claude Code outright, so the cheapest route in is a paid subscription or API credentials. HIPAA coverage reaches only Enterprise and the API and depends on a Business Associate Agreement that Anthropic grants case by case after review, and Claude Team is explicitly outside that scope. Claude Code is not the right fit for developers who want inline autocomplete rather than conversational task execution. GitHub Copilot starts at $10/mo and Cursor at $16/mo billed annually, and both put suggestions directly in the editor without a separate terminal workflow. It is also not an open-source project, despite a public repository with a large following. That repository carries the plugins, changelog and issue tracker under a license reading All rights reserved, with use governed by Anthropic's Commercial Terms of Service, so the agent itself cannot be forked, self-hosted or independently audited. Current state Q3 2026: the public repository has 143k stars and more than 22,600 forks, with commits landing daily. Fable 5 access was suspended on June 12 under US government export controls and restored on July 1 once those controls were lifted, and it is available in Claude Code. Sonnet 5 launched on June 30, 2026 and Opus 5 on July 24, 2026. Anthropic states that inputs and outputs from its commercial products, including the API and Claude for Work, are not used for model training by default. Consumer Pro and Max accounts instead carry a model training setting the user can switch off to stop new chats and coding sessions being used. The Claude platform holds SOC 2 Type II, ISO 27001:2022 and ISO/IEC 42001:2023, with CSA STAR, NIST 800-171 and UK Cyber Essentials also published, and FedRAMP High scoped to the separate Claude for Government offering.

Pricing

subscription · $17/mo annual

View pricing ↗

Segment

both

Setup

moderate

Verified

Aug 12, 2026

Transparency

Mostly Public

Contract

Monthly or Annual

Data training

Opt-out

Human in loop

Optional

Capabilities

terminal-agentagentic-codinglarge-codebaseautonomous

Pros & Limitations

Editorial assessment

Pros

  • Available across every developer surface from one subscription: terminal, VS Code, JetBrains, Slack, web, desktop, iOS and GitHub Actions, with no model differences between them, which removes the context-switching cost of single-surface tools.
  • Works as an MCP client and an MCP server at once. It connects out to external servers over HTTP, SSE, stdio and WebSocket, and the claude mcp serve command exposes Claude Code itself for other applications to connect into.
  • Runs on Anthropic's own frontier models rather than a third-party API, so new releases reach it immediately. Fable 5, Opus 5, Sonnet 5 and Haiku 4.5 are all selectable, at published per-token rates when billed through the API.

Limitations

  • There is no standalone Claude Code plan and no free tier. It is bundled into Claude subscriptions and shares one usage pool with chats, so terminal work draws down the same rolling five-hour and weekly limits, and the Free plan excludes it entirely.
  • It is not open source, despite the public GitHub repository. That repository holds plugins, the changelog and the issue tracker under a license reading All rights reserved, with use governed by Anthropic's Commercial Terms of Service, so the agent cannot be forked, self-hosted or independently audited.
  • HIPAA coverage is limited to Enterprise and the API and requires a Business Associate Agreement granted case by case after review. Claude Team is explicitly out of scope, and consumer Pro and Max accounts rely on a user-managed model training setting rather than a contractual commitment.

Technical Details

Deployment
cliidewebdesktopmobile
Model architectureProprietary (Claude Fable, Opus, Sonnet and Haiku)
Autonomous rateDeveloper-supervised by default: the agent proposes changes and runs commands with review at each step, and an opt-in auto mode removes the per-step prompt for long-running work.
Integrations
VS CodeJetBrainsGitHubGitLabGitHub ActionsSlack
Security
SOC 2 Type IIISO 27001ISO 42001HIPAAGDPRCCPA

Similar agents

Rating

4.6/ 5

Editorial score

How we score this →

Recognition

Rated 4.6 ★Category Leader · Q3 2026MCP Server VerifiedListed 2026
Embed these on your site →

Industries

EnterpriseDevToolsSaaSEducation

Leave a review

Never displayed publicly.

Agent Stacks

See workflow stacks that feature Claude Code.

Compare

Related Content

Is this your tool?

Claim this listing free to get your embeddable badges, a review collection link, and the Verified checkmark.

Claim listing →

Free · Every Two Weeks

AI Agent Price & Rating Tracker

Price changes, new agent launches, acquisitions, and rating updates across the AI agent market. Verified against live vendor data, not vendor marketing.

No spam. Unsubscribe anytime. We never share your email.