Cursor

Cursor is an AI-first code editor designed around two complementary superpowers:

1. Tab autocomplete for staying in flow on the line you’re writing
2. Agent workflows for when the task is bigger than a single file (refactors, fixes, scaffolds, “make this work end-to-end”)

If you’re building with tools like Codex, Claude Code, or Gemini and you want them living inside your editor with awareness of your repo, Cursor is one of the most common “vibe coder to pro dev” bridges.

Pricing Tiers

Cursor offers individual and business plans. The pricing page is the source of truth, but the gist is: Free for a taste, Pro for real daily use, Pro+/Ultra for heavy frontier-model usage, and Teams/Enterprise for governance.

How usage works (important if you’re budgeting)

Cursor’s Pro plan includes unlimited Tab and unlimited model usage when Auto is selected, plus a monthly pool of included usage for frontier models at API pricing with an option to buy more and set spend limits. (Their blog post explains the mechanics and why they moved away from request-based limits.)

Core Features & Capabilities

• Agent mode (the “do the task” assistant)
  Cursor’s Agent can tackle multi-step work, run terminal commands, and edit code as it goes, so it can reproduce issues and validate fixes instead of only suggesting snippets.

• Tab completion (the “stay in flow” assist)
  Cursor keeps the classic IDE loop fast with inline suggestions, which is often where most of the day-to-day value lives.

• Model flexibility
  Cursor supports frontier coding models from major providers and lets you choose models explicitly or use Auto routing depending on your plan and needs.

• Team & enterprise controls
  The Teams/Enterprise plans add identity management, provisioning, role controls, privacy mode controls, reporting, and admin governance to keep agentic tooling usable in real orgs.

• Security posture + privacy controls
  Cursor documents how it approaches AI requests, codebase indexing, and privacy mode, and advertises SOC 2 certification and trust resources for teams.

Pros & Cons

Pros:

• A “single cockpit” for both autocomplete and agentic work.
• Lots of model choice, which helps you match the tool to the task (fast edits vs deep refactors).
• Strong options for teams (SSO, RBAC, admin/model controls).

Cons:

• Heavy frontier-model usage can turn into real money; watch included usage, spend limits, and what you select vs Auto.
• Agents can make big changes quickly, so you want a simple ritual: small tasks, clear acceptance criteria, review diffs, commit often.
• Like most AI coding tools, output quality improves dramatically when your repo has clear structure and you provide “rules of the road.”

Summary

Further reading

	cursor.com	Cursor pricing
	cursor.com	Agent overview (terminal + code edits)
	cursor.com	Supported models
	cursor.com	How Pro usage-based pricing works
	cursor.com	Security (privacy mode, codebase indexing, AI requests)
	cursor.com	Enterprise & team controls
	trust.cursor.com	Trust Center