AI Agent Index
ByHeather MacAvelia·Independently reviewed·Published Mar 18, 2026·Updated Sep 7, 2026
Independently verified against live vendor data on Sep 7, 2026.

AI-first code editor with autonomous agent mode, cloud subagents, self-hosted runners and Origin code hosting. Free tier, with Individual from $16/month billed annually (USD).

How we scored it

Autonomy

4/5

Integrations

5/5

Pricing clarity

4/5

Evidence

5/5

Setup

5/5

The facts

Cursor is an AI-first code editor built by Anysphere as a fork of VS Code, with AI embedded at every layer of the coding workflow. On June 16, 2026, SpaceX announced a definitive agreement to acquire Anysphere for $60 billion in an all-stock transaction. That acquisition has not closed: Cursor's terms of service and its site footer both name Anysphere, Inc. as the contracting entity. Cursor states that 64% of Fortune 500 companies use the product, that more than 50,000 enterprises build with it, and that over 100 million lines of enterprise code are written with it per day. Cursor indexes entire repositories, reads across all files, and writes directly to the codebase. It documents fourteen first-party integrations: GitHub, GitLab, Bitbucket, Azure DevOps, Jira, Linear, Slack, Microsoft Teams, Notion, JetBrains, Xcode, Google Drive, Gmail and Google Calendar. A public Marketplace adds installable plugins, automations and skills, with a publish path for teams building their own. Cursor is an MCP client rather than an MCP server: it connects out to external MCP servers configured through mcp.json or the Customize page, and Teams administrators can distribute MCP servers across cloud agents, the IDE and the CLI. Cursor publishes no MCP server of its own, re-verified September 2026. The Cursor SDK and a Cloud Agents API let teams build programmatic agents deployable from CI/CD pipelines and internal tools. All prices are USD. Hobby is free with no credit card, limited agent requests and limited tab completions. Individual is $16 per month billed annually or $20 month to month, with Pro, Pro+ and Ultra sub-tiers at $20, $60 and $200 month to month for progressively higher model limits. Teams has two tiers: Standard at $32 per user per month billed annually or $40 month to month, and Premium at $96 billed annually or $120 month to month. Enterprise is quote only. Annual billing is 20% below the month-to-month rate on every tier. The entry Individual tier does not buy centralized billing, team-wide privacy mode, SAML/OIDC SSO, usage analytics, agentic code review included at no extra charge, or the team marketplace for internal rules, skills and plugins. All of those start at Teams, and Bugbot runs on usage-based billing on Individual. Repository, model and MCP access controls, SCIM seat management, audit logs, the AI code tracking API and pooled usage are Enterprise only. HIPAA Business Associate Agreements and US-only data residency are Enterprise only too, and data residency adds a 10% uplift on model pricing for eligible models. Every plan includes a set amount of model usage, and on-demand usage continues past that allowance billed in arrears. Invoicing and wire transfer are Enterprise only, and self-serve plans take cards. Cursor suits engineering teams that want an AI-native editor as their primary environment and are comfortable with model inference running in Cursor's cloud. Teams wanting fully on-premise AI coding with local model hosting should look at Sourcegraph Cody (custom pricing), which supports air-gapped deployments. Teams invested in JetBrains IDEs can keep the native environment with JetBrains AI Assistant (from $8.33/mo billed annually, on top of an existing JetBrains license). For terminal-first, git-native workflows without an IDE layer, Claude Code (from $17/mo billed annually) offers deeper git integration. GitHub Copilot (from $10/mo) is the tighter fit for teams that do not want to leave GitHub's own tooling. Current state Q3 2026: Origin, Cursor's own code hosting, entered early beta on all paid plans on August 17, 2026, with repos, pull requests, code browsing and GitHub sync. For Origin-hosted repos, Origin is the source of truth rather than GitHub. Self-Hosted Machines shipped on September 2, 2026, moving Cloud Agent tool execution onto hardware the customer manages, with pooled workers and hibernation. Cloud Agents no longer require a connected source-control provider. Composer 2.5 is Cursor's proprietary model, selectable alongside Grok 4.6, Grok 4.5, GPT-5.6, Claude Opus 5, Claude Sonnet 5, Claude Fable 5.1 and Gemini 3.1 Pro. Grok Bot is included on Individual and Teams plans and carries its own terms. Security: Cursor holds ISO/IEC 27001:2022, ISO/IEC 42001:2023 and AIUC-1 certifications alongside a SOC 2 Type II attestation, with certificates available on request through its trust center. The enterprise page states compliance with the requirements of GDPR and CCPA, HIPAA Business Associate Agreements are available on the Enterprise plan, and Cursor maintains zero data retention agreements with its model providers.

Pricing

freemium · $16/mo annual

View pricing ↗

Segment

both

Setup

easy

Verified

Sep 7, 2026

Transparency

Mostly Public

Contract

Monthly or Annual

Data training

Opt-out

Human in loop

Optional

Capabilities

idemulti-file-editingautocompleteagentic-codingmobileautomations

Pros & Limitations

Editorial assessment

Pros

  • Agent mode plans and implements features autonomously across multiple files, and cloud subagents run in parallel on their own VMs and branches.
  • Full VS Code compatibility means near-zero migration cost, because existing extensions, keybindings, themes and workflows carry over immediately.
  • Privacy Mode is available on every plan including the free tier, and is on by default for Teams and Enterprise, where admins can enforce it.
  • Self-Hosted Machines runs Cloud Agent tool execution on hardware the team manages, keeping the codebase, build outputs and secrets on internal machines.

Limitations

  • Privacy Mode is off by default on the Hobby and Individual plans, where Cursor states it may store and use codebase data, prompts, editor actions and code snippets to improve the product.
  • Model inference always routes through Cursor's cloud, including when you supply your own API keys, and zero data retention does not apply to those keys. US-only data residency is Enterprise only.
  • Expensive model usage depletes included credits faster than standard models, and agentic code review with Bugbot runs on usage-based billing on Individual rather than being included as it is on Teams.

Technical Details

Deployment
desktopclicloudmobile
Model architectureMulti-model (SpaceXAI, OpenAI, Anthropic, Google Gemini) plus proprietary Composer
Autonomous rateComposer agent autonomously plans and implements features across multiple files. Version 3.6 Auto-review Run Mode enables longer autonomous runs with fewer approval prompts. Version 3.7 adds cloud subagents that work on their own VMs and branches in parallel. Developer reviews the diff before committing: human-in-the-loop by design.
Integrations
GitHubGitLabBitbucketAzure DevOpsJiraLinearSlackMicrosoft TeamsNotionJetBrainsXcodeGoogle DriveGmailGoogle Calendar
Security
SOC 2 Type IIISO 27001ISO 42001GDPRCCPAHIPAAAIUC-1

Similar agents

Rating

4.6/ 5

Editorial score

How we score this →

Recognition

Rated 4.6 ★Category Leader · Q3 2026Listed 2026
Embed these on your site →

Industries

FinanceEcommerceMedia

Built for

SaaSStartupsEnterpriseDevTools

Leave a review

Never displayed publicly.

Agent Stacks

See workflow stacks that feature Cursor.

Compare

Related Content

Is this your tool?

Claim this listing free to get your embeddable badges, a review collection link, and the Claimed badge on your listing.

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.